Multi-omics evaluation

Abstract

Described herein are methods, such as multi-omics methods, for assessing diseases, such as cancer. The multi-omics methods can integrate proteomics data, transcriptomics data, genomics data, lipidomics data, or metabolomics data. Methods for screening for diseases or conditions. Also described herein are methods for screening for diseases or conditions from biological samples. The methods can include assessing whether a nodule, mass, or cyst is cancerous.

Description

[Technical Field]

[0001] cross reference This application is a continuation of U.S. Provisional Application No. 63 / 168,594, filed March 31, 2021; U.S. Provisional Application No. 63 / 168,634, filed March 31, 2021; U.S. Provisional Application No. 63 / 183,816, filed May 4, 2021; U.S. Provisional Application No. 63 / 183,829, filed May 4, 2021; Provisional Application No. 63 / 183,844, filed May 4, 2021, U.S. Provisional Application No. 63 / 183,852, filed May 5, 2021, U.S. Provisional Application No. 63 / 184,498, filed May 5, 2021, U.S. Provisional Application No. 63 / 228,533, filed August 2, 2021, U.S. Provisional Application No. 63 / 228,543, filed August 2, 2021, U.S. Provisional Application No. 63 / 228,544, filed August 2, 2021, This application claims the benefit of U.S. Provisional Application No. 63 / 229,232 filed on August 4, 2021, U.S. Provisional Application No. 63 / 229,242 filed on August 4, 2021, U.S. Provisional Application No. 63 / 256,482 filed on October 15, 2021, U.S. Provisional Application No. 63 / 278,637 filed on November 12, 2021, U.S. Provisional Application No. 63 / 288,825 filed on December 13, 2021, U.S. Provisional Application No. 63 / 288,827 filed on December 13, 2021, U.S. Provisional Application No. 63 / 312,455 filed on February 22, 2022, and U.S. Provisional Application No. 63 / 322,149 filed on March 21, 2022, each of which is incorporated herein by reference in its entirety. [Background technology]

[0002] There is a need for methods for the early and accurate detection of disease conditions such as cancer. Accurate and early disease detection can improve treatment and prognosis for diseased subjects. Summary of the Invention [Means for solving the problem]

[0003] In some embodiments, disclosed herein are multi-omics methods, including obtaining multi-omics data generated from one or more biological fluid samples collected from a subject suspected of having a disease state, the multi-omics data including proteomic measurements and nucleic acid sequencing measurements; applying a classifier to the multi-omics data to assess the disease state; and any one of (i) to (iv): (i) the proteomic measurements are generated after samples of the one or more biological fluid samples are subjected to an enrichment protocol that enriches for a protein or peptide without enriching for another protein or peptide; (ii) the proteomic measurements are generated after samples of the one or more biological fluid samples are subjected to an enrichment protocol that enriches for a protein or peptide without enriching for another protein or peptide; (iii) the classifier comprises a performance characteristic comprising a mean or median area under the curve (AUC) of at least 0.9, determined in a dataset derived from a randomized controlled trial of at least 20 subjects with the disease state and more than 20 control subjects without the disease state; or (iv) the evaluation comprises selecting a cancer therapy based on the multi-omics data, and the proteomic measurements are generated using mass spectrometry. In some embodiments, the proteomic measurements are generated after a sample of one or more biofluid samples undergoes an enrichment protocol that enriches some proteins without enriching other proteins. In some embodiments, the proteomic measurements are generated from proteins adsorbed to nanoparticles. In some embodiments, the proteomic measurements are generated based on the amount of protein added to a sample of one or more biofluid samples. In some embodiments, the proteins added to the sample are labeled. In some embodiments, the nucleic acid sequencing measurements include mRNA sequencing measurements. In some embodiments, the nucleic acid sequencing measurements include mRNA sequencing measurements and miRNA sequencing measurements. In some embodiments, the multi-omics data includes measurements of more than 45 peptide or protein groups.In some embodiments, the performance of the assessment is at least 4% better than when the classifier is applied to only one type of omics data, and the performance includes sensitivity at a given specificity as determined in a dataset derived from a randomized controlled trial of more than 25 subjects with the condition and more than 25 control subjects without the condition. In some embodiments, the classifier is characterized by a mean area under the receiver operating characteristic (ROC) curve (AUC) of at least 0.9 as determined in a dataset derived from a randomized controlled trial of at least 20 subjects with the condition and more than 20 control subjects without the condition. In some embodiments, applying a classifier to the multi-omic data to assess a disease state includes applying a first classifier to the proteomic measurements to generate a first label corresponding to the presence, absence, or likelihood of the disease state, applying a second classifier to the nucleic acid sequencing measurements to generate a second label corresponding to the presence, absence, or likelihood of the disease state, and assessing the disease state based on (a), (b), or (c): (a) an unweighted average of the first and second labels, (b) a weighted average of the first and second labels, or (c) a majority score based on the first and second labels. Some embodiments include assessing the disease state based on a weighted average of the first and second labels, wherein the weighted average is generated by assigning weights to the results of the first and second classifiers based on an area under the receiver operating characteristic curve (ROC) curve, an area under the precision-recall curve, precision, precision, recall, sensitivity, F1-score, specificity, or a combination thereof. In some embodiments, applying the classifier to the multi-omics data to assess the disease state includes obtaining a subset of features from the proteomic measurements; obtaining at least a subset of features from the nucleic acid sequencing measurements; pooling the subset of features from the first omics data and the at least a subset of features from the second omics data into obtained pooled features; and assessing the disease state based on the pooled features. In some embodiments, obtaining the subset of features from the first or second omics data includes obtaining top features based on univariate data.In some embodiments, the classifier is trained using deep learning, hierarchical cluster analysis, principal component analysis, partial least squares discriminant analysis, random forest classification analysis, support vector machine analysis, k-nearest neighbor analysis, naive Bayes analysis, k-means clustering analysis, or hidden Markov analysis. In some embodiments, the multi-omics data further comprises metabolomics data. In some embodiments, the disease state comprises cancer. In some embodiments, the cancer is selected from the group consisting of lung cancer, pancreatic cancer, breast cancer, colon cancer, liver cancer, and ovarian cancer. In some embodiments, the assessment comprises selecting a cancer therapy based on the multi-omics data. In some embodiments, the assessment comprises administering a chemotherapy, pharmaceutical, radiation, or surgical cancer treatment to the subject based on the assessment. In some embodiments, the one or more biological fluid samples comprise a blood, serum, or plasma sample. In some embodiments, the subject is human. Provided herein, in some embodiments, is a method for obtaining multi-omics data generated from one or more blood, serum, or plasma samples taken from a human subject suspected of having cancer, the multi-omics data including proteomic measurements and RNA sequencing measurements; applying a classifier to the multi-omics data to assess the cancer; selecting or administering a cancer therapy to the subject based on the assessment; and any one of (i)-(iii): (i) the proteomic measurements are generated after samples of the one or more blood, serum, or plasma samples are enriched with affinity reagents for proteins or peptides; (ii) the proteomic measurements are generated based on the amount of labeled proteins or peptides added to the one or more blood, serum, or plasma samples; or (iii) the classifier is generated using hold-out data derived from a randomized controlled trial of at least 25 subjects with the condition and more than 25 control subjects without the condition. a performance characteristic comprising a mean area under the curve (AUC) of a receiver operating characteristic (ROC) curve of at least 0.9 as determined on a set of (data).In some embodiments, the proteomic measurements are generated after one or more blood, serum, or plasma samples are enriched with an affinity reagent. In some embodiments, the proteomic measurements are generated based on the amount of labeled protein added to the one or more blood, serum, or plasma samples. In some embodiments, the classifier is characterized by a mean area under the curve (AUC) of a receiver operating characteristic (ROC) curve of at least 0.9, as determined in a dataset derived from a randomized controlled trial of at least 25 subjects with the condition and more than 25 control subjects without the condition.

[0004] Disclosed herein, in some aspects, is a multi-omic disease detection method comprising: acquiring multi-omic data generated from one or more biological fluid samples collected from a subject, the multi-omic data comprising first omics data comprising proteomics data, metabolomics data, transcriptomics data, or genomics data, and second omics data comprising proteomics data, metabolomics data, transcriptomics data, or genomics data different from the first omics data; and assigning, using a first classifier, a first label to the first omics data comprising the presence, absence, or likelihood of a disease state; assigning, using a second classifier, a second label to the second omics data comprising the presence, absence, or likelihood of a disease state based on the first and second labels; and identifying the multi-omic data as indicative or not indicative of a disease state. In some embodiments, the first omics data includes proteomics data, and the second omics data includes metabolomics data, transcriptomics data, or genomics data. In some embodiments, the proteomics data is generated by contacting one of the biofluid samples with particles so that the particles adsorb biomolecules, including proteins. In some embodiments, the particles include carboxylate particles, polyacrylic acid particles, dextran particles, polystyrene particles, dimethylamine particles, amino particles, silica particles, or N-(3-trimethoxysilylpropyl)diethylenetriamine particles. In some embodiments, the particles include physiochemically distinct nanoparticles. In some embodiments, the proteomics data is generated using mass spectrometry, chromatography, liquid chromatography, high-performance liquid chromatography, solid-phase chromatography, lateral flow assay, immunoassay, enzyme-linked immunosorbent assay, Western blot, dot blot, or immunostaining, or a combination thereof.In some embodiments, the genomics data or transcriptomics data is generated by sequencing, microarray analysis, hybridization, polymerase chain reaction, electrophoresis, or a combination thereof. In some embodiments, the second omics data comprises transcriptomics data. In some embodiments, the transcriptomics data comprises mRNA or microRNA expression data. In some embodiments, the second omics data comprises genomics data. In some embodiments, the genomics data comprises DNA sequence data or epigenetic data. In some embodiments, identifying the multi-omics data as indicative of or not indicative of a disease state comprises identifying the multi-omics data as indicative of or not indicative of a disease state based on either the first label or the second label. In some embodiments, identifying the multi-omics data as indicative of or not indicative of a disease state comprises generating or obtaining a majority score based on the first and second labels. In some embodiments, identifying the multi-omics data as indicative or non-indicative of a disease state includes generating or obtaining a weighted average of the first and second labels. Some embodiments include assigning weights to the first and second classifiers based on an area under a receiver operating characteristic (ROC) curve, an area under a precision-recall curve, accuracy, precision, recall, sensitivity, F1-score, specificity, or a combination thereof, thereby obtaining the weighted average. In some embodiments, the first omics data is generated from a first one of the biological fluid samples, and the second omics data is generated from a second one of the biological fluid samples. In some embodiments, the first biological fluid sample is collected in a first container containing a first collection component including heparin, ethylenediaminetetraacetic acid (EDTA), citrate, or an anti-lytic agent, and the second biological fluid sample is collected in a second container containing a second collection component different from the first collection component, the second collection component including heparin, EDTA, citrate, or an anti-lytic agent. In some embodiments, the multi-omics data further comprises third omics data comprising a third omics data type.The third omics data may comprise a different omics data type or subtype than the first and second omics data. Some embodiments include using a third classifier to assign a third label to the third omics data corresponding to the presence, absence, or likelihood of a disease state. In some embodiments, identifying the multi-omics data as indicative of or not indicative of a disease state comprises identifying the multi-omics data as indicative of or not indicative of a disease state based on a combination of the first, second, and third labels. Some embodiments include using a third classifier to assign a third label comprising the presence, absence, or likelihood of a disease state to the third omics data different from the first and second omics data, and identifying the multi-omics data as indicative of or not indicative of a disease state based on the first and second labels comprises identifying the multi-omics data as indicative of or not indicative of a disease state based on the first, second, and third labels. In some embodiments, the first omics data type comprises proteomics data, the second omics data type comprises mRNA transcriptomics data, and the third omics data type comprises microRNA transcriptomics data. Some embodiments include transmitting or outputting information related to the identification. Some embodiments include recommending a treatment for the condition.

[0005] Disclosed herein, in some embodiments, is a method comprising obtaining combined data comprising two, three, or four of proteomic, metabolomic, transcriptomic, or genomic data generated from one or more biological fluid samples from a subject; and using a classifier to classify the combined data as indicative or not indicative of one or more disease states. In some embodiments, the one or more biological fluid samples comprise 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, or more biological fluid samples. In some embodiments, the combined data is generated simultaneously. In some embodiments, the simultaneous data generation comprises simultaneously assaying two, three, or four of the proteomic, metabolomic, transcriptomic, or genomic data. In some embodiments, the simultaneous data generation comprises assaying two, three, or four of the proteomic, metabolomic, transcriptomic, or genomic data at separate locations on an assay substrate. In some embodiments, the distinct locations comprise distinct wells, and the assay substrate comprises an assay plate. In some embodiments, the one or more biological fluid samples comprise two or more of a whole blood sample, a plasma sample, a serum sample, or a urine sample. In some embodiments, proteomic data is generated from one of the one or more biological fluid samples. In some embodiments, metabolomic data is generated from the biological fluid sample or from an additional biological fluid sample of the one or more biological fluid samples, and the proteomic data and metabolomic data are combined to obtain combined data. In some embodiments, the classifier identifies the combined data as indicative or non-indicative of one or more disease states with greater sensitivity or specificity than the proteomic data, metabolomic data, transcriptomic data, or genomic data alone. In some embodiments, the classifier includes features selected from the proteomic data, metabolomic data, genomic data, or transcriptomic data.In some embodiments, the classifier comprises features selected from a combination of proteomic data, metabolomic data, genomic data, or transcriptomic data. In some embodiments, the classifier comprises multiple classifiers. In some embodiments, the multiple classifiers comprise two, three, four, or more classifiers. In some embodiments, the multiple classifiers separately comprise features selected from proteomic data, metabolomic data, genomic data, transcriptomic data, or a combination thereof. In some embodiments, using a classifier to identify the combined data as indicative of or not indicative of one or more disease states comprises using multiple classifiers to identify the combined data as indicative of or not indicative of one or more disease states. In some embodiments, using a classifier to identify the combined data as indicative of or not indicative of one or more disease states comprises selecting the output of any one of the multiple classifiers. In some embodiments, using a classifier to identify the combined data as indicative of or not indicative of one or more disease states comprises majority voting among the multiple classifiers. In some embodiments, using a classifier to identify the combined data as indicative of or not indicative of one or more disease states comprises a majority vote across a subset of a plurality of classifiers. In some embodiments, using a classifier to identify the combined data as indicative of or not indicative of one or more disease states comprises a weighted average of a plurality of classifiers. In some embodiments, using a classifier to identify the combined data as indicative of or not indicative of one or more disease states comprises a weighted average of a subset of a plurality of classifiers. In some embodiments, the weights of the weighted average are assigned based on an area under a receiver operating characteristic (ROC) curve. In some embodiments, the weights of the weighted average are assigned based on an area under a precision-recall curve. In some embodiments, the weights of the weighted average are assigned based on accuracy. In some embodiments, the weights of the weighted average are assigned based on precision. In some embodiments, the weights of the weighted average are assigned based on recall. In some embodiments, the weights of the weighted average are assigned based on sensitivity.In some embodiments, the weights in the weighted average are assigned based on the F1-score. In some embodiments, the weights in the weighted average are assigned based on the specificity.

[0006] Disclosed herein, in some embodiments, are methods that include obtaining proteomic data generated from a biological fluid sample from a subject; obtaining metabolomic, transcriptomic, or genomic data generated from the biological fluid sample from the subject or from an additional biological fluid sample, combining the proteomic data with the metabolomic, transcriptomic, or genomic data to obtain combined data; and using a classifier to identify the combined data as indicative or non-indicative of one or more pathological conditions. In some embodiments, the proteomic data is generated by contacting the biological fluid sample from the subject with particles such that the particles adsorb biomolecules, including proteins. Some embodiments include contacting the biological fluid sample from the subject with particles such that the particles adsorb the biomolecules. Some embodiments include analyzing the biomolecules adsorbed to the particles to generate the proteomic data. Some embodiments include analyzing the biological fluid sample or the additional biological fluid sample to generate the metabolomic data. Some embodiments include using a classifier to identify the combined data as indicative or non-indicative of one or more pathologies. In some embodiments, the proteomic data is generated by measuring readouts indicative of the presence, absence, or quantity of biomolecules. In some embodiments, the proteomic data is generated using mass spectrometry, chromatography, liquid chromatography, high-performance liquid chromatography, solid-phase chromatography, lateral flow assays, immunoassays, enzyme-linked immunosorbent assays, Western blots, dot blots, or immunostaining, or a combination thereof. In some embodiments, the proteomic data is generated using mass spectrometry. In some embodiments, the protein comprises a secreted protein. In some embodiments, the particle comprises a nanoparticle. In some embodiments, the particle comprises a lipid particle, a metal particle, a silica particle, or a polymer particle.In some embodiments, the particles comprise carboxylate particles, polyacrylic acid particles, dextran particles, polystyrene particles, dimethylamine particles, amino particles, silica particles, or N-(3-trimethoxysilylpropyl)diethylenetriamine particles. In some embodiments, the particles comprise physiochemically distinct nanoparticle populations. In some embodiments, the metabolomics data is generated from a different biological fluid sample than the proteomics data. In some embodiments, the metabolomics data is generated using mass spectrometry, electrophoresis, colorimetric assay, fluorescence assay, chromatography, liquid chromatography, high-performance liquid chromatography, solid-phase chromatography, lateral flow assay, immunoassay, or a combination thereof. In some embodiments, the metabolomics data is generated using mass spectrometry. In some embodiments, the metabolomics data is generated from the same biological fluid sample as the proteomics data. In some embodiments, the metabolomics data is generated by analyzing analytes adsorbed to the particles. In some embodiments, the metabolomics data comprises lipid metabolite data, carbohydrate metabolite data, vitamin metabolite data, cofactor metabolite data, or a combination thereof. In some embodiments, the biological fluid sample comprises a blood sample, a plasma sample, or a serum sample. In some embodiments, the additional biological fluid sample is collected from the subject in a container separate from the biological fluid sample. In some embodiments, the combined data is generated from 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, or more samples. In some embodiments, the 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, or more samples are collected separately in 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, or more containers. In some embodiments, the 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, or more containers contain multiple components in addition to the sample. In some embodiments, the biological fluid sample and the additional biological fluid sample are collected in separate containers, and these separate containers contain different components. In some embodiments, a first one of the separate containers contains a first component that is different from a second component in a second one of the separate containers.In some embodiments, the biological fluid sample comprises serum; is collected in a container comprising ethylenediaminetetraacetic acid (EDTA), citrate, or heparin; or comprises a preservative to prevent cell lysis. In some embodiments, the biological fluid sample is collected in a container comprising ethylenediaminetetraacetic acid (EDTA). In some embodiments, the additional biological fluid sample comprises a blood sample, a plasma sample, or a serum sample. In some embodiments, the additional biological fluid sample has been processed to obtain cell-free DNA or to obtain RNA. Some embodiments include obtaining genomics or transcriptomics data generated from the biological fluid sample from the subject, from the additional biological fluid sample, or from a third biological fluid sample. In some embodiments, the combined data further comprises genomics or transcriptomics data. Some embodiments include analyzing the biological fluid sample, the additional biological fluid sample, or the third biological fluid sample to generate genomics or transcriptomics data. In some embodiments, the third biological fluid sample comprises a blood sample, a plasma sample, or a serum sample. In some embodiments, the third biological fluid sample is processed to obtain cell-free DNA or RNA. Some embodiments include using a classifier to identify the combined data as indicating or not indicating one or more pathologies. In some embodiments, the genomics data or transcriptomics data is generated by measuring readouts indicating the presence, absence, or amount of nucleic acids. In some embodiments, the genomics data or transcriptomics data is generated by sequencing, microarray analysis, hybridization, polymerase chain reaction, electrophoresis, or a combination thereof. In some embodiments, the genomics data or transcriptomics data is generated from a different biological fluid sample than the metabolomics data. In some embodiments, the genomics data or transcriptomics data is generated from the same biological fluid sample as the metabolomics data.In some embodiments, the genomics data or transcriptomics data is generated from a different biofluid sample than the proteomics data. In some embodiments, the genomics data or transcriptomics data is generated from the same biofluid sample as the proteomics data. In some embodiments, the genomics data or transcriptomics data is generated by analyzing nucleic acids adsorbed to particles. In some embodiments, the genomics data or transcriptomics data comprises genomics data. In some embodiments, the genomics data comprises DNA sequence data. In some embodiments, the genomics data comprises DNA polymorphism data. In some embodiments, the genomics data comprises epigenetic data. In some embodiments, the genomics data comprises DNA methylation data. In some embodiments, the epigenetic data comprises histone modification data. In some embodiments, the histone modification data comprises acetylation data, methylation data, ubiquitination data, phosphorylation data, sumoylation data, ribosylation data, or citrullination data. In some embodiments, the genomics data or transcriptomics data comprises transcriptomics data. In some embodiments, the transcriptomics data comprises RNA sequence data. In some embodiments, the transcriptomics data comprises RNA expression data. In some embodiments, the transcriptomics data comprises mRNA, tRNA, rRNA, microRNA, snRNA, snoRNA, or lncRNA expression data. In some embodiments, the transcriptomics data comprises mRNA expression data. In some embodiments, the transcriptomics data comprises microRNA expression data. In some embodiments, the classifier comprises features for identifying the combined data as indicative of one or more disease states. In some embodiments, the features comprise control protein measurements, control metabolite measurements, control nucleic acid measurements, mass spectra, m / z ratios, chromatography results, immunoassay results, light or fluorescence intensity, or sequence information.In some embodiments, the classifier is trained using deep learning, hierarchical cluster analysis, principal component analysis, partial least squares discriminant analysis, random forest classification analysis, support vector machine analysis, k-nearest neighbor analysis, naive Bayes analysis, k-means clustering analysis, or hidden Markov analysis. In some embodiments, the one or more disease states include one or more cancers. In some embodiments, the one or more cancers include lung cancer, breast cancer, prostate cancer, colorectal cancer, colon cancer, melanoma, bladder cancer, lymphoma, leukemia, kidney cancer, uterine cancer, pancreatic cancer, or a combination thereof. In some embodiments, the classifier distinguishes between one or more disease states. In some embodiments, the classifier distinguishes between lung cancer, colon cancer, and pancreatic cancer. In some embodiments, the classifier distinguishes between lung cancer, colon cancer, and pancreatic cancer. In some embodiments, the lung cancer includes non-small cell lung cancer (NSCLC). Some embodiments include generating a report based on use of the classifier to identify the combined data as indicative or not indicative of one or more disease states. In some embodiments, the report includes a likelihood or indication that the biological fluid or the subject contains one or more disease states. Some embodiments include outputting or transmitting the report. In some embodiments, the report is used by a medical professional in making a diagnosis, providing medical advice, or providing treatment for at least one of the one or more disease states. Some embodiments include identifying the combined data as indicative of one or more disease states. In some embodiments, the one or more disease states include cancer, and if the combined data is identified as indicative of cancer, further including recommending a cancer treatment to the subject. In some embodiments, the one or more disease states include cancer, and if the combined data is identified as indicative of cancer, further including administering a cancer treatment to the subject. In some embodiments, the cancer treatment includes chemotherapy, radiation therapy, ablation therapy, embolization, or surgery. Some embodiments include using a classifier to identify the combined data as indicative of a first of the one or more disease states and not indicative of a second of the one or more disease states. Some embodiments involve administering or recommending treatment for the first condition but not the second condition.Some embodiments include identifying the combined data as not indicative of one or more pathological conditions. Some embodiments include observing the subject without providing treatment to the subject if the combined data is identified as not indicative of one or more pathological conditions. In some embodiments, observing the subject without providing treatment includes analyzing biomolecules in a biological fluid sample subsequently obtained from the subject. In some embodiments, the subject is a mammal. In some embodiments, the subject is human. In some embodiments, the classifier includes features selected from proteomics data, metabolomics data, genomics data, or transcriptomics data. In some embodiments, the classifier includes features selected from proteomics data, metabolomics data, genomics data, or transcriptomics data. In some embodiments, the classifier comprises a plurality of classifiers. In some embodiments, the plurality of classifiers comprises two, three, or four or more classifiers. In some embodiments, the plurality of classifiers separately comprise features selected from proteomics data, metabolomics data, genomics data, transcriptomics data, or combinations thereof. In some embodiments, using a classifier to identify the combined data as indicative of or not indicative of one or more disease states comprises using a plurality of classifiers to identify the combined data as indicative of or not indicative of one or more disease states. In some embodiments, using a classifier to identify the combined data as indicative of or not indicative of one or more disease states comprises selecting the output of any one of the plurality of classifiers. In some embodiments, using a classifier to identify the combined data as indicative of or not indicative of one or more disease states comprises majority voting among the plurality of classifiers. In some embodiments, using a classifier to identify the combined data as indicative of or not indicative of one or more disease states comprises majority voting among a subset of the plurality of classifiers. In some embodiments, using a classifier to classify the combined data as indicative or not indicative of one or more disease states comprises a weighted average of a plurality of classifiers. In some embodiments, using a classifier to classify the combined data as indicative or not indicative of one or more disease states comprises a weighted average of a subset of the plurality of classifiers. In some embodiments, the weights of the weighted average are assigned based on an area under a receiver operating characteristic (ROC) curve. In some embodiments, the weights of the weighted average are assigned based on an area under a precision-recall curve. In some embodiments, the weights of the weighted average are assigned based on accuracy. In some embodiments, the weights of the weighted average are assigned based on precision. In some embodiments, the weights of the weighted average are assigned based on recall. In some embodiments, the weights of the weighted average are assigned based on sensitivity. In some embodiments, the weights of the weighted average are assigned based on F1-score. In some embodiments, the weights of the weighted average are assigned based on specificity.

[0007]

[0003] Disclosed herein, in some aspects, is a method comprising: acquiring multi-omics data generated from one or more biological fluid samples collected from a subject, the multi-omics data including first omics data and second omics data, wherein the first omics data includes a first omics data type including proteomics data, metabolomics data, transcriptomics data, or genomics data, and the second omics data includes a second omics data type different from the first omics data type, including proteomics data, metabolomics data, transcriptomics data, or genomics data; identifying a first subset of features from the first omics data; identifying a second subset of features from the second omics data; pooling the first and second subsets of features; and identifying the multi-omics data as indicative or not indicative of a disease state based on the pooled subset of features. In some embodiments, identifying the first or second subset of features from the first or second omics data includes obtaining univariate data for features of the first or second omics data and identifying the first or second subset based on the univariate data. In some embodiments, the first or second subset of features is identified from features of a classifier for the first or second omics data. In some embodiments, identifying the first or second subset of features from the first or second omics data includes obtaining a classifier for the first or second omics data and identifying the first or second subset as top features for the classifier. In some embodiments, identifying the first or second subset of features from the first or second omics data includes obtaining a classifier for the first or second omics data, removing one or more features at a time from the classifier, and identifying which features, when removed from the classifier, degrade the performance of the classifier.

[0008] In some embodiments, the disease or disorder comprises pancreatic cancer. In some aspects, a multi-omic cancer detection method for detecting pancreatic cancer is disclosed herein. In some aspects, a method for detecting pancreatic cancer in a subject is disclosed herein, the method comprising: identifying a subject at risk for pancreatic cancer; obtaining a biological fluid sample from the subject; contacting the biological fluid sample with particles such that the particles adsorb biomolecules, including proteins, to the particles; assaying the biomolecules adsorbed to the particles to generate proteomic data; and classifying the proteomic data as indicative of pancreatic cancer or not indicative of pancreatic cancer. In some aspects, a method is disclosed herein, the method comprising: assaying proteins in a biological fluid sample obtained from a subject identified as at risk for pancreatic cancer to obtain protein measurements; and applying a classifier to the protein measurements to thereby identify the protein measurements as indicative of the subject having pancreatic cancer, wherein the classifier is generated using proteomic data obtained by contacting a training sample with particles such that the particles adsorb proteins in the training sample and assaying the proteins adsorbed to the particles. Disclosed herein, in some aspects, are methods of treatment that include identifying a mass in the pancreas of a subject; obtaining a biological fluid sample from the subject; contacting the biological fluid sample with particles such that the particles adsorb biological molecules, including proteins, to the particles; assaying the biological molecules adsorbed to the particles to generate proteomic data; and classifying the proteomic data as indicative of a mass comprising pancreatic cancer or as not indicative of a mass comprising pancreatic cancer.Provided herein, in some embodiments, is a method of evaluating a subject suspected of having pancreatic cancer, comprising measuring biomarkers in a biological fluid sample from the subject, the biomarkers being A2GL, AKR1B1, ANPEP, ANTXR1, ANTXR2, BTK, CALR, CDH1, CDH11, CDH2, CDHR2, CILP2, CLEC3B, COL18A1, CRP, EXT1, F13A1, FAT1, FGL1, FLT 4, ICAM1, IDH2, LCN2, LPP, MAPK1, MAP2K1, MYH9, NOTCH1, NOTCH2, PIGR, PPP2R1A, PRKAR1A, PXDN, RELN, RHOA, S100A8, S100A9, S100A12, SAA1, SAA2, SERPINA3, SLAIN2, SND1, SVEP1, TSP2, TUBB, TUBB1, or VCAN. In some embodiments, a method is disclosed herein that includes assaying biomolecules in a biological fluid sample obtained from a subject suspected of having pancreatic cancer to obtain biomolecule measurements; and applying a classifier to the biomolecule measurements to identify protein measurements as indicative of the subject having or not having pancreatic cancer, wherein the classifier is characterized by a receiver operating characteristic (ROC) curve based on the biomolecule measurement features, having an area under the curve (AUC) of greater than 0.7, greater than 0.75, greater than 0.8, greater than 0.85, greater than 0.9, greater than 0.91, greater than 0.92, greater than 0.93, or greater than 0.94. In some embodiments, the AUC is 0.75 or less, 0.8 or less, 0.85 or less, 0.9 or less, 0.91 or less, 0.92 or less, 0.93 or less, 0.94 or less, 0.95 or less, or 0.96 or less. In some embodiments, the biomolecules include proteins, lipids, or metabolites, or a combination thereof.

[0009] In some embodiments, the disease or disorder comprises liver cancer. In some aspects, a multi-omic cancer detection method for detecting liver cancer is disclosed herein. In some aspects, a method for detecting liver cancer in a subject is disclosed herein, comprising: identifying the subject as being at risk for liver cancer; obtaining a biological fluid sample from the subject; contacting the biological fluid sample with particles such that the particles adsorb biomolecules, including proteins, to the particles; assaying the biomolecules adsorbed to the particles to generate proteomic data; and classifying the proteomic data as indicating liver cancer or not indicating liver cancer. In some aspects, a method is disclosed herein, comprising: assaying proteins in a biological fluid sample obtained from a subject identified as being at risk for liver cancer to obtain protein measurements; and applying a classifier to the protein measurements to thereby identify the protein measurements as indicating that the subject has liver cancer, wherein the classifier is generated using proteomic data obtained by contacting a training sample with particles such that the particles adsorb proteins in the training sample and assaying the proteins adsorbed to the particles. In some embodiments, disclosed herein is a method for treatment, comprising: identifying a mass in the liver of an object; obtaining a biological fluid sample from the object; contacting the biological fluid sample with particles, so that the particles adsorb biomolecules, including proteins, to the particles; assaying the biomolecules adsorbed to the particles to generate proteomic data; and classifying the proteomic data as indicating a mass containing liver cancer or as not indicating liver cancer.In some embodiments, disclosed herein is a method for detecting liver cancer in an object, comprising: identifying the object as having a risk of liver cancer; obtaining a biological fluid sample from the object; assaying lipids in the biological fluid sample to obtain lipid data; and classifying the lipid data as indicating liver cancer or as not indicating liver cancer.

[0010] In some embodiments, the disease or disorder comprises ovarian cancer. In some aspects, a multi-omic cancer detection method for detecting ovarian cancer is disclosed herein. In some aspects, a method for detecting ovarian cancer in a subject is disclosed herein, the method comprising: identifying the subject as being at risk for ovarian cancer; obtaining a biological fluid sample from the subject; contacting the biological fluid sample with particles such that the particles adsorb biomolecules, including proteins, to the particles; assaying the biomolecules adsorbed to the particles to generate proteomic data; and classifying the proteomic data as being indicative of ovarian cancer or not indicative of ovarian cancer. In some aspects, identifying the subject as being at risk for ovarian cancer comprises identifying the subject as having undergone a computed tomography (CT) scan indicative of ovarian cancer, a magnetic resonance imaging (MRI) scan indicative of ovarian cancer, a positron emission tomography (PET) scan indicative of ovarian cancer, a transvaginal ultrasound indicative of ovarian cancer, an elevated cancer antigen (CA)-125 level compared to a control or baseline measurement, or an ovarian cyst, or a combination thereof. Disclosed herein, in some embodiments, is a method comprising: assaying proteins in a biological fluid sample obtained from a subject identified as being at risk for ovarian cancer to obtain protein measurements; and applying a classifier to the protein measurements to thereby identify the protein measurements as indicative of the subject having ovarian cancer, wherein the classifier is generated using proteomic data obtained by contacting a training sample with particles such that the particles adsorb proteins in the training sample and assaying the proteins adsorbed to the particles. In some embodiments, the proteins include ANTXR2, BMP1, CILP, EIF2AK2, ENO3, F13B, FGL1, or PEBP4.In some embodiments, disclosed herein is a method for treatment, comprising: identifying a mass in the ovary of a subject; obtaining a biological fluid sample from the subject; contacting the biological fluid sample with particles, so that the particles adsorb biomolecules, including proteins, to the particles; assaying the biomolecules adsorbed to the particles to generate proteomic data; and classifying the proteomic data as indicating a mass containing ovarian cancer or as not indicating ovarian cancer.In some embodiments, disclosed herein is a method for detecting ovarian cancer in a subject, comprising: identifying the subject as having a risk of having ovarian cancer; obtaining a biological fluid sample from the subject; assaying lipids in the biological fluid sample to obtain lipid data; and classifying the lipid data as indicating ovarian cancer or as not indicating ovarian cancer.In some embodiments, the lipid comprises one or more phospholipids.

[0011] In some embodiments, the disease or disorder comprises colorectal cancer. In some aspects, disclosed herein are multi-omic cancer detection methods for detecting colorectal cancer. In some aspects, disclosed herein are methods for detecting colorectal cancer in a subject, the methods comprising: identifying the subject as at risk for having colorectal cancer; obtaining a biological fluid sample from the subject; contacting the biological fluid sample with particles such that the particles adsorb biomolecules, including proteins, to the particles; assaying the biomolecules adsorbed to the particles to generate proteomic data; and classifying the proteomic data as indicative of colorectal cancer or not indicative of colorectal cancer. In some aspects, disclosed herein are methods comprising: assaying proteins in a biological fluid sample obtained from a subject identified as at risk for colorectal cancer to obtain protein measurements; and applying a classifier to the protein measurements to thereby identify the protein measurements as indicative of the subject having colorectal cancer, wherein the classifier is generated using proteomic data obtained by contacting a training sample with particles such that the particles adsorb proteins in the training sample and assaying the proteins adsorbed to the particles. In some embodiments, a subject is identified as being at risk for having colorectal cancer by identifying the subject as having undergone a computed tomography (CT) scan indicative of colorectal cancer, undergoing liver function tests (LFTs) indicative of colorectal cancer, having elevated carcinoembryonic antigen (CEA) levels compared to a control or baseline measurement, having blood in the stool, undergoing a fecal immunochemical test (FIT) indicative of colorectal cancer, or having colon nodules, or a combination thereof. Disclosed herein, in some embodiments, are methods of treatment that include identifying a mass in the subject's colon; obtaining a biological fluid sample from the subject; contacting the biological fluid sample with particles such that the particles adsorb biomolecules, including proteins, to the particles; assaying the biomolecules adsorbed to the particles to generate proteomic data; and classifying the proteomic data as indicative of a mass comprising colorectal cancer or as not indicative of colorectal cancer.

[0012] In some embodiments, the present disclosure provides a method for assessing the lung nodule by assaying proteins in a biological fluid sample obtained from a subject identified as having a lung nodule to obtain protein measurements; and applying a classifier to the protein measurements to assess the lung nodule; and (i), (ii), or (iii): (i) the classifier comprises protein features of the assayed proteins, and the classifier comprises performance characteristics in distinguishing the lung nodule as cancerous or non-cancerous, determined on a dataset derived from a randomized controlled trial of more than 20 subjects with cancerous lung nodules and more than 20 control subjects with non-cancerous lung nodules, without including clinical features in the classifier. Disclosed methods include: (i) the classifier comprises performance characteristics determined on a dataset, the performance characteristics comprising a mean or median area under the curve (AUC) of a receiver operating characteristic (ROC) curve greater than 0.65 (e.g., greater than 0.7); (ii) the classifier is generated using proteomic data obtained by contacting a training sample with particles such that the particles adsorb proteins in the training sample and assaying the proteins adsorbed to the particles; or (iii) assaying the proteins comprises contacting a biological fluid sample with particles to adsorb proteins to the particles and obtaining protein measurements from the adsorbed proteins. In some embodiments, the classifier is characterized by a mean ROC curve that includes protein features of the assayed proteins and has a median AUC of greater than 0.7 in classifying lung nodules as cancerous or non-cancerous, the AUC greater than 0.7 being determined without including non-protein features on a dataset derived from a randomized controlled trial of greater than 20 subjects with cancerous lung nodules and greater than 20 control subjects with non-cancerous lung nodules. In some embodiments, the classifier is generated using proteomic data obtained by contacting a training sample with particles such that the particles adsorb proteins in the training sample and assaying the proteins adsorbed to the particles, hi some embodiments, assaying the proteins comprises contacting a biological fluid sample with the particles to adsorb proteins to the particles and obtaining protein measurements from the adsorbed proteins.In some embodiments, the classifier is trained using deep learning, hierarchical cluster analysis, principal component analysis, partial least squares discriminant analysis, random forest classification analysis, support vector machine analysis, k-nearest neighbor analysis, naive Bayes analysis, k-means clustering analysis, or hidden Markov analysis. In some embodiments, assessing the lung nodule includes identifying protein measurements as indicative of the lung nodule being cancerous. Some embodiments include administering a lung cancer treatment to the subject based on the assessment. In some embodiments, the lung cancer treatment includes chemotherapy, radiation therapy, percutaneous ablation, radiofrequency ablation, cryoablation, microwave ablation, chemoembolization, or surgery. In some embodiments, the subject is identified as having a lung nodule through the use of a medical imaging device. In some embodiments, the classifier identifies lung cancer with a sensitivity and specificity greater than 60%. In some embodiments, the particles include nanoparticles. In some embodiments, the particles include lipid particles, metal particles, silica particles, or polymer particles. In some embodiments, the particles include physiochemically distinct nanoparticles. In some embodiments, the biological fluid sample comprises a blood, serum, or plasma sample. In some embodiments, the subject is a human. In some embodiments, the protein measurements comprise measurements of a protein selected from the group consisting of APP, IGHG2, SERPING1, SAA2, SERPINF2, GC, IGHA1, HPR, SERPINA3, IGHA1, LTF, SERPINA1, PCSK6, PROS1, BPIF1, C6, CP, A2M, and IGFBP2.Provided herein, in some aspects, is a method for assaying proteins in a blood, serum, or plasma sample by mass spectrometry to obtain protein measurements, wherein the sample is obtained from a human subject identified as having a lung nodule using a medical imaging device; applying a classifier to the protein measurements to assess the lung nodule; and selecting or administering a lung cancer therapy to the subject based on the assessment; and (i), (ii), or (iii): (i) the classifier comprises protein features of the assayed proteins, and the classifier comprises a performance characteristic in identifying the lung nodule as cancerous or non-cancerous, wherein the classifier is selected from more than 25 subjects with cancerous lung nodules and more than 25 controls with non-cancerous lung nodules. Disclosed are methods including: (i) the classifier comprises performance characteristics comprising a median area under the curve (AUC) of a receiver operating characteristic (ROC) curve greater than 0.7, determined on a holdout dataset derived from a randomized controlled trial of the subject, and determined using only protein features of the classifier; (ii) the classifier is generated using proteomic data obtained by contacting training samples with nanoparticles such that the nanoparticles adsorb proteins in the training samples and assaying the proteins adsorbed to the nanoparticles; or (iii) assaying the proteins comprises contacting a blood, serum, or plasma sample with nanoparticles to adsorb proteins to the nanoparticles, and obtaining protein measurements from the adsorbed proteins.

[0013] In some embodiments, the classifier is characterized by an average ROC curve that includes protein features of the assayed proteins and has a median AUC of greater than 0.7 in classifying lung nodules as cancerous or non-cancerous, where the AUC of greater than 0.7 is determined without including non-protein features in a holdout dataset derived from a randomized controlled trial of greater than 25 subjects with cancerous lung nodules and greater than 25 control subjects with non-cancerous lung nodules. In some embodiments, the classifier is generated using proteomic data obtained by contacting training samples with nanoparticles such that the nanoparticles adsorb proteins in the training samples and assaying the proteins adsorbed to the nanoparticles. In some embodiments, assaying the proteins includes contacting a blood, serum, or plasma sample with nanoparticles to adsorb proteins to the nanoparticles and obtaining protein measurements from the adsorbed proteins.

[0014] Disclosed herein in some embodiments are methods comprising: assaying proteins in a biological fluid sample obtained from a subject identified as having a lung nodule to obtain protein measurements; and applying a classifier to the protein measurements to identify the protein measurements as indicative of the lung nodule being cancerous or non-cancerous, wherein the classifier is characterized by a receiver operating characteristic (ROC) curve based on the protein measurement features having an area under the curve (AUC) of greater than 0.7. In some embodiments, the AUC greater than 0.7 is generated without including non-protein clinical features. In some embodiments, the non-protein clinical features include clinical indicators of lung cancer. In some embodiments, the proteins include APP, IGHG2, SERPING1, SAA2, SERPINF2, GC, IGHA1, HPR, SERPINA3, IGHA1, LTF, SERPINA1, PCSK6, PROS1, BPIF1, C6, CP, A2M, or IGFBP2.

[0015] Disclosed herein, in some embodiments, is a method comprising: assaying proteins in a biological fluid sample obtained from a subject having or suspected of having a lung nodule to obtain protein measurements; and applying a classifier to the protein measurements to assess the lung nodule, the classifier being generated using proteomic data obtained by enriching proteins with an affinity reagent. Disclosed herein, in some embodiments, is a method comprising: assaying proteins in a biological fluid sample obtained from a subject having or suspected of having a lung nodule to obtain protein measurements; and applying a classifier to the protein measurements to thereby identify the protein measurements as indicative of the lung nodule being cancerous or non-cancerous, the classifier being generated using proteomic data obtained by contacting a training sample with particles such that the particles adsorb proteins in the training sample and assaying the proteins adsorbed to the particles. Some embodiments include obtaining or receiving a biological fluid sample from a subject. In some embodiments, the subject is identified as having a lung nodule by medical imaging. In some embodiments, the medical imaging includes a computed tomography (CT) scan. Some embodiments include performing medical imaging. Some embodiments include identifying a lung nodule in the medical imaging. Some embodiments include generating a report based on identifying the protein measurements indicating whether the lung nodule is cancerous or non-cancerous. In some embodiments, the report includes a likelihood or indication that the lung nodule is cancerous or non-cancerous. Some embodiments include outputting or transmitting the report. In some embodiments, the report is used by a medical professional in making a diagnosis, providing medical advice, or providing treatment for the lung nodule. Some embodiments include performing a biopsy on the lung nodule if the protein measurements are classified as indicating the lung nodule is cancerous. In some embodiments, the biopsy confirms the likelihood that the lung nodule is cancerous or non-cancerous. In some embodiments, the lung nodule is cancerous.In some embodiments, the lung nodule comprises non-small cell lung cancer (NSCLC). In some embodiments, the classifier comprises features indicative of protein measurements that indicate whether the lung nodule is cancerous or non-cancerous. In some embodiments, the features comprise control protein measurements, mass spectra, m / z ratios, chromatography results, immunoassay results, or light or fluorescence intensity. In some embodiments, the classifier is trained using deep learning, hierarchical cluster analysis, principal component analysis, partial least squares discriminant analysis, random forest classification analysis, support vector machine analysis, k-nearest neighbor analysis, naive Bayes analysis, k-means clustering analysis, or hidden Markov analysis. In some embodiments, the classifier can identify lung cancer with a sensitivity of 50% or greater, 60% or greater, 70% or greater, 80% or greater, or 90% or greater. In some embodiments, the classifier can identify lung cancer with a specificity of 50% or greater, 60% or greater, 70% or greater, 80% or greater, or 90% or greater. Some embodiments include recommending a lung cancer treatment to the subject if the protein measurements are classified as indicating that the lung nodule is cancerous. Some embodiments include administering a lung cancer treatment to the subject if the protein measurements are classified as indicating that the lung nodule is cancerous. In some embodiments, the lung cancer treatment includes chemotherapy, radiation therapy, percutaneous ablation, radiofrequency ablation, cryoablation, microwave ablation, chemoembolization, or surgery. In some embodiments, the lung nodule is non-cancerous. Some embodiments include observing the subject without performing a biopsy if the protein measurements are classified as indicating that the lung nodule is non-cancerous. In some embodiments, observing the subject without performing a biopsy includes assaying for proteins in a second biological fluid sample later obtained from the subject. Some embodiments include assaying for proteins in a second biological fluid sample later obtained from the subject. In some embodiments, the particles include nanoparticles. In some embodiments, the particles include lipid particles, metal particles, silica particles, or polymer particles.In some embodiments, the particles comprise carboxylate particles, polyacrylic acid particles, dextran particles, polystyrene particles, dimethylamine particles, amino particles, silica particles, or N-(3-trimethoxysilylpropyl)diethylenetriamine particles. In some embodiments, the particles comprise physiochemically distinct nanoparticles. In some embodiments, assaying for protein comprises contacting a biological fluid sample with particles such that the particles adsorb the protein to the particles. In some embodiments, assaying for protein comprises measuring a readout indicative of the presence, absence, or amount of the biomolecule. In some embodiments, assaying for protein comprises performing mass spectrometry, chromatography, liquid chromatography, high-performance liquid chromatography, solid-phase chromatography, lateral flow assay, immunoassay, enzyme-linked immunosorbent assay, Western blot, dot blot, or immunostaining, or a combination thereof. In some embodiments, assaying for protein comprises performing mass spectrometry. In some embodiments, the protein comprises a secreted protein. In some embodiments, the biological fluid comprises blood, plasma, or serum. In some embodiments, the lung nodule is less than 3 cm in diameter. In some embodiments, the subject has multiple lung nodules. In some embodiments, the subject is a mammal. In some embodiments, the subject is human.

[0016] In some embodiments, a method is disclosed herein that includes obtaining a biological fluid sample from a subject having a lung nodule; contacting the biological fluid sample with particles such that the particles adsorb biomolecules, including proteins, to the particles; assaying the biomolecules adsorbed to the particles to generate proteomic data; and classifying the proteomic data as indicative of whether the lung nodule is cancerous or non-cancerous. In some embodiments, the subject is identified as having a lung nodule by medical imaging. In some embodiments, the medical imaging includes a computed tomography (CT) scan. Some embodiments include performing the medical imaging. Some embodiments include identifying the lung nodule in the medical imaging. Some embodiments include performing a biopsy on the lung nodule if the proteomic data is classified as indicative of the lung nodule being cancerous. In some embodiments, the biopsy confirms the likelihood that the lung nodule is cancerous or non-cancerous. In some embodiments, the lung nodule is cancerous and comprises a tumor. In some embodiments, the lung nodule comprises non-small cell lung cancer (NSCLC). In some embodiments, classifying the proteomic data as indicative of whether the lung nodule is cancerous or non-cancerous includes applying a classifier to the proteomic data. In some embodiments, the classifier includes features indicative of the likelihood that the lung cancer is cancerous or non-cancerous. In some embodiments, the classifier is trained using deep learning, hierarchical cluster analysis, principal component analysis, partial least squares discriminant analysis, random forest classification analysis, support vector machine analysis, k-nearest neighbor analysis, naive Bayes analysis, k-means clustering analysis, or hidden Markov analysis. In some embodiments, the proteomic data indicates that the lung nodule is cancerous or non-cancerous with a sensitivity or specificity of about 80% or greater. Some embodiments include recommending a lung cancer treatment to the subject if the proteomic data is classified as indicative of the lung nodule being cancerous. Some embodiments include administering a lung cancer treatment to the subject if the proteomic data is classified as indicative of the lung nodule being cancerous.In some embodiments, the lung cancer treatment comprises chemotherapy, radiation therapy, percutaneous ablation, radiofrequency ablation, cryoablation, microwave ablation, chemoembolization, or surgery. In some embodiments, the lung nodule is non-cancerous and benign. Some embodiments include observing the subject without biopsy if the proteomic data classifies the lung nodule as being non-cancerous. Some embodiments include monitoring the subject and subsequently assaying biomolecules in a second biological fluid sample obtained from the subject. In some embodiments, the particles comprise nanoparticles. In some embodiments, the particles comprise lipid particles, metal particles, silica particles, or polymer particles. In some embodiments, the particles comprise carboxylate particles, polyacrylic acid particles, dextran particles, polystyrene particles, dimethylamine particles, amino particles, silica particles, or N-(3-trimethoxysilylpropyl)diethylenetriamine particles. In some embodiments, the particles comprise physiochemically distinct nanoparticle populations. In some embodiments, assaying the biomolecule comprises measuring a readout indicative of the presence, absence, or amount of the biomolecule. In some embodiments, assaying the biomolecule comprises performing mass spectrometry, chromatography, liquid chromatography, high performance liquid chromatography, solid phase chromatography, lateral flow assay, immunoassay, enzyme-linked immunosorbent assay, Western blot, dot blot, or immunostaining, or a combination thereof. In some embodiments, assaying the biomolecule comprises performing mass spectrometry. In some embodiments, the protein comprises a secreted protein. In some embodiments, the biological fluid comprises blood, plasma, or serum. In some embodiments, the lung nodule is less than 3 cm in diameter. In some embodiments, the subject has multiple lung nodules. In some embodiments, the subject is a mammal. In some embodiments, the subject is a human.

[0017] Disclosed herein in some embodiments is a method comprising: assaying proteins in a biological fluid sample obtained from a subject suspected of having a pulmonary nodule to obtain protein measurements; and applying a classifier to the protein measurements to thereby identify the protein measurements as indicative of the subject having a pulmonary nodule, the classifier being generated using proteomic data obtained by contacting a training sample with particles such that the particles adsorb proteins in the training sample and assaying the proteins adsorbed to the particles. Some embodiments include recommending that the subject undergo medical imaging, such as a CT scan, if the protein measurements indicate that the subject has a pulmonary nodule; and not recommending that the subject undergo medical imaging if the protein measurements do not indicate that the subject has a pulmonary nodule. Some embodiments include performing medical imaging, such as a CT scan, on the subject if the protein measurements indicate that the subject has a pulmonary nodule; and not performing medical imaging on the subject if the protein measurements do not indicate that the subject has a pulmonary nodule. Some embodiments include sending or receiving a report on medical imaging, such as a CT scan, if the protein measurements indicate the subject has a lung nodule, and not sending or receiving a report if the protein measurements do not indicate the subject has a lung nodule. In some embodiments, the protein measurements indicate that the subject has or is likely to have a lung nodule. In some embodiments, the protein measurements indicate that the subject does not have or is unlikely to have a lung nodule.

[0018] In some embodiments, a method is disclosed herein that includes assaying proteins in a biological fluid sample obtained from a subject suspected of having lung cancer to obtain protein measurements; and applying a classifier to the protein measurements to thereby identify the protein measurements as indicative of the subject having lung cancer, where the classifier is generated using proteomic data obtained by contacting a training sample with particles such that the particles adsorb proteins in the training sample and assaying the proteins adsorbed to the particles. Some embodiments include recommending that the subject undergo medical imaging, such as a CT scan, if the protein measurements indicate that the subject has lung cancer; and not recommending that the subject undergo medical imaging if the protein measurements do not indicate that the subject has lung cancer. Some embodiments include performing medical imaging, such as a CT scan, on the subject if the protein measurements indicate that the subject has lung cancer; and not performing medical imaging on the subject if the protein measurements do not indicate that the subject has lung cancer. Some embodiments include sending or receiving a report regarding medical imaging, such as a CT scan, if the protein measurements indicate that the subject has lung cancer; and not sending or receiving a report if the protein measurements do not indicate that the subject has lung cancer. In some embodiments, the protein measurement indicates that the subject has or is likely to have lung cancer. In some embodiments, the protein measurement indicates that the subject does not have or is unlikely to have lung cancer. In some embodiments, lung cancer includes NSCLC.

[0019] In some embodiments, a method is disclosed herein that includes obtaining a biological fluid sample from a subject suspected of having a pulmonary nodule; contacting the biological fluid sample with particles such that the particles adsorb biomolecules, including proteins, to the particles; assaying the biomolecules adsorbed to the particles to generate proteomic data; and classifying the proteomic data based on the proteomic data as indicating that the subject has a pulmonary nodule or as not indicating that the subject has a pulmonary nodule. Some embodiments include recommending that the subject undergo medical imaging, such as a CT scan, if the proteomic data indicates that the subject has a pulmonary nodule; and not recommending that the subject undergo medical imaging if the proteomic data does not indicate that the subject has a pulmonary nodule. Some embodiments include performing medical imaging, such as a CT scan, on the subject if the proteomic data indicates that the subject has a pulmonary nodule; and not performing medical imaging on the subject if the proteomic data does not indicate that the subject has a pulmonary nodule. Some embodiments include sending or receiving a report on medical imaging, such as a CT scan, if the proteomic data indicates the subject has a lung nodule, and not sending or receiving a report if the proteomic data does not indicate the subject has a lung nodule. In some embodiments, the proteomic data indicates that the subject has or is likely to have a lung nodule. In some embodiments, the proteomic data indicates that the subject does not have or is unlikely to have a lung nodule.

[0020] In some embodiments, a method is disclosed herein that includes obtaining a biological fluid sample from a subject suspected of having lung cancer; contacting the biological fluid sample with particles so that the particles adsorb biomolecules, including proteins, to the particles; assaying the biomolecules adsorbed to the particles to generate proteomic data; and classifying the proteomic data based on the proteomic data as indicating that the subject has lung cancer or as not indicating that the subject has lung cancer. Some embodiments include recommending that the subject undergo medical imaging, such as a CT scan, if the proteomic data indicates that the subject has lung cancer, and not recommending that the subject undergo medical imaging if the proteomic data does not indicate that the subject has lung cancer. Some embodiments include performing medical imaging, such as a CT scan, on the subject if the proteomic data indicates that the subject has lung cancer, and not performing medical imaging on the subject if the proteomic data does not indicate that the subject has lung cancer. Some embodiments include sending or receiving a report on medical imaging, such as a CT scan, if the proteomic data indicates that the subject has lung cancer, and not sending or receiving a report if the proteomic data does not indicate that the subject has lung cancer. In some embodiments, the proteomic data indicates that the subject has or is likely to have lung cancer. In some embodiments, the proteomic data indicates that the subject does not have or is unlikely to have lung cancer.

[0021] In some embodiments, a monitoring method is disclosed herein that includes obtaining a biological fluid sample from a subject at risk of lung cancer recurrence; contacting the biological fluid sample with particles such that the particles adsorb biomolecules, including proteins, to the particles; assaying the biomolecules adsorbed to the particles to generate proteomic data; and classifying the proteomic data based on the proteomic data as indicating that the subject has lung cancer recurrence or as not indicating that the subject has lung cancer recurrence. Some embodiments include recommending that the subject undergo medical imaging, such as a CT scan, if the protein measurements indicate that the subject has lung cancer recurrence, and not recommending that the subject undergo medical imaging if the protein measurements do not indicate that the subject has lung cancer recurrence. Some embodiments include performing medical imaging, such as a CT scan, on the subject if the protein measurements indicate that the subject has lung cancer recurrence, and not performing medical imaging on the subject if the protein measurements do not indicate that the subject has lung cancer recurrence. Some embodiments include sending or receiving a report on medical imaging, such as a CT scan, if the protein measurements indicate the subject has lung cancer recurrence, and not sending or receiving a report if the protein measurements do not indicate the subject has lung cancer recurrence. In some embodiments, the protein measurements indicate the subject has or is likely to have lung cancer recurrence. In some embodiments, the protein measurements indicate the subject does not have or is unlikely to have lung cancer recurrence. In some embodiments, the subject is undergoing lung cancer treatment. In some embodiments, the lung cancer treatment includes chemotherapy, radiation therapy, or surgery. In some embodiments, the cancer is potentially resectable. In some embodiments, the lung cancer includes NSCLC. [Brief explanation of the drawings]

[0022] [Figure 1A] A multi-omics approach is shown.

[0023] [Figure 1B]1 shows the combination of datasets in a multi-omics approach.

[0024] [Figure 2A] 1 illustrates an example method for generating and applying a classifier as described herein.

[0025] [Figure 2B] 1 is a flow chart illustrating some aspects that may be used in the methods herein.

[0026] [Figure 3A] 1 illustrates exemplary steps in the screening and treatment of patients with or suspected of having a pathological condition.

[0027] [Figure 3B] 1 illustrates exemplary steps in the screening and treatment of patients with pancreatic cancer.

[0028] [Figure 3C] 1 shows examples of steps in the screening and treatment of liver cancer patients.

[0029] [Figure 4] A non-limiting example of a computing device is shown (in this case, a device with one or more processors, memory, storage, and network interfaces).

[0030] [Figure 5] FIG. 1 illustrates a diagram of a classifier and feature information according to some aspects described herein.

[0031] [Figure 6] 1 shows a graph illustrating the differential expression of several proteins that can be used to generate a classifier for diagnosing disease states.

[0032] [Figure 7]Figure 1 shows a diagram showing the expression of several proteins in samples from diseased subjects compared to control subjects. Several genes were differentially expressed (under- or over-expressed) between the groups (NSCLC samples and healthy samples).

[0033] [Figure 8] Shown are scatter plots of pairwise predictions plotted against each other. RNASeq: Predicted probability based on RNA-Seq data (affected). Proteomics: Predicted probability based on proteomics data (affected). RNA_Prot: Predicted probability based on both RNA-Seq and proteomics data (affected).

[0034] [Figure 9] Receiver operating characteristic (ROC) curves are included, showing the increase in area under the curve (AUC) when combining mRNA transcriptomics and proteomics data compared to mRNA transcriptomics or proteomics data alone.

[0035] [Figure 10A] shows an additive multi-omics classification of 30 samples from subjects with a pathological condition and 30 samples from control subjects, including mRNA transcriptomics data, proteomics data, and a combination of mRNA transcriptomics and proteomics data.

[0036] [Figure 10B] 1 shows the differential mRNAs and proteins whose abundances were measured in the biofluid samples and used to generate the classifiers.

[0037] [Figure 11A] Analyses based on proteomic and microRNA data are shown. The top panel shows the results of a classifier trained on proteomic data only, the middle panel shows the results of a classifier trained on microRNA data only, and the bottom panel shows the results of combining the two data types.

[0038] [Figure 11B] The differentially expressed microRNAs used to generate the classifier are shown.

[0039] [Figure 12] An analysis comparing the combination of three omics data types (proteomics, mRNA, and miRNA) versus using only one of each of the three data types is shown.

[0040] [Figure 13A] 1 illustrates several aspects that can be used for integrated model classification.

[0041] [Figure 13B] 1 illustrates several aspects that can be used in transformation-based classification.

[0042] [Figure 14] 1 shows the graphical results of the integrated model classification analysis.

[0043] [Figure 15] 1 illustrates some aspects of a transformation-based classification analysis.

[0044] [Figure 16] 1 shows the graphical results of the integrated model classification analysis and transformation-based classification.

[0045] [Figure 17] 1 shows a non-limiting example of a flow chart of a machine training algorithm for improving the sensitivity and specificity of a classifier for predicting disease as described herein.

[0046] [Figure 18A] 1 shows ROC curves for several protein data and combined protein + lipid data for pathology classification.

[0047] [Figure 18B]This includes aspects of the sensitivity of analysis of protein data, lipid data, and combined protein+lipid data for disease state classification.

[0048] [Figure 19] 1 illustrates aspects of a two-stage machine learning framework for analyzing and training multiple data types.

[0049] [Figure 20A] This includes aspects of the sensitivity of analysis of protein data, lipid data, and combined protein+lipid data for disease state classification.

[0050] [Figure 20B] This includes aspects of the sensitivity of analysis of protein data, lipid data, and combined protein+lipid data for disease state classification.

[0051] [Figure 20C] The ROC curves for several protein data, lipid data, and combined protein and lipid data for pathological classification are shown.

[0052] [Figure 21] The ROC curves for several protein data sets for disease state classification and combined protein, lipid, and clinical parameter data sets are shown.

[0053] [Figure 22A] Some protein data information is shown.

[0054] [Figure 22B] Performance aspects of several classifiers are shown.

[0055] [Figure 22C] We show the performance aspects of several classifiers with and without several features.

[0056] [Figure 23]Some genetic or transcriptional data aspects are shown, such as the index or type of measurement, the type of sample, quality control aspects, or sequencing depth that can be used.

[0057] [Figure 24] Illustrated are various embodiments that can be used in some of the methods described herein.

[0058] [Figure 25] The methods described herein include several aspects such as subjects or test results that can be included.

[0059] [Figure 26A] It includes a table showing some proteins, their OT scores, and a description of some features in the protein classifier.

[0060] [Figure 26B] It includes a table showing some proteins, their OT scores, and a description of some features in the protein classifier.

[0061] [Figure 27] Includes a chart showing feature importance scores for the lipid classifier.

[0062] [Figure 28A] The results of the Wilcox test for age comparisons and Fisher's exact test for gender ratios are shown.

[0063] [Figure 28B] The results of the Wilcox test for age comparisons and Fisher's exact test for gender ratios are shown.

[0064] [Figure 29A] The number of proteins detected across samples of interest in the analysis of biofluid samples from control and cancer patients is shown.

[0065] [Figure 29B] The number of proteins detected across samples of interest in the analysis of biofluid samples from control and cancer patients is shown.

[0066] [Figure 30A] A plot of some of the top proteins differentially detected in biofluid samples from cancer patients compared to biofluid samples from control patients is shown.

[0067] [Figure 30B] This is a plot showing the distribution of OpenTargets (OT) scores. OT scores (0-0.8) are on the x-axis, and density (0-15) is on the y-axis.

[0068] [Figure 31A] Included are plots showing a comparison of median total signal by sample, analyte type, and class.

[0069] [Figure 31B] 1 shows box plots of the most significantly different analytes per omics workflow; top left: lipids; bottom left: metabolites; and right: proteins), according to one embodiment.

[0070] [Figure 31C] An example of the performance of a multi-omic classifier combining proteomic, lipidomic, and metabolomic measurements is shown.

[0071] [Figure 32A] Included is a volcano plot of the intensity difference and P-value of proteins adsorbed to nanoparticles and detected in biofluid samples from cancer patients compared to biofluid samples from control patients, with the magnitude of the difference displayed on the x-axis and significance displayed on the y-axis, highlighting the most significant analytes.

[0072] [Figure 32B]Contains data for the top protein P35442 after particle-based measurement methods.

[0073] [Figure 32C] Volcano plots of the intensity differences and P-values ​​of proteins detected in biofluid samples from cancer patients compared to biofluid samples from control patients are included, with the magnitude of the difference displayed on the x-axis and significance displayed on the y-axis, highlighting the most significant analytes.

[0074] [Figure 32D] Contains data for the top protein P01011 after proteomics measurements.

[0075] [Figure 33A] Included is a volcano plot of the intensity differences and P-values ​​of lipids detected in biofluid samples from cancer patients compared to biofluid samples from control patients, with the magnitude of the difference displayed on the x-axis and significance displayed on the y-axis, highlighting the most significant analytes.

[0076] [Figure 33B] Includes data on the top lipid CER (d18:1_18:0) after lipidomics measurement.

[0077] [Figure 34A] Volcano plots of the intensity differences and P-values ​​of metabolites detected in biofluid samples from cancer patients compared to biofluid samples from control patients are included, with the magnitude of the difference displayed on the x-axis and significance displayed on the y-axis, highlighting the most significant analytes.

[0078] [Figure 34B] Includes data on the top metabolite AICAR after metabolomics measurements.

[0079] [Figure 35A] 1 shows the classification of cancer and healthy samples by UMAP projection based on the combined data.

[0080] [Figure 35B] 1 shows the classification of cancer and healthy samples by PCA projection based on the combined data.

[0081] [Figure 35C] Classification of cancer and healthy samples by UMAP projection based on Proteograph data.

[0082] [Figure 35D] 1 shows the classification of cancer and healthy samples by PCA projection based on Proteograph data.

[0083] [Figure 35E] Classification of cancer and healthy samples by UMAP projection based on PiQuant data.

[0084] [Figure 35F] Classification of cancer and healthy samples by PCA projection based on PiQuant data.

[0085] [Figure 35G] 1 shows the classification of cancer and healthy samples by UMAP projection based on lipid data.

[0086] [Figure 35H] 1 shows the classification of cancer and healthy samples by PCA projection based on lipid data.

[0087] [Figure 35I] 1 shows the classification of cancer and healthy samples by UMAP projection based on metabolite data.

[0088] [Figure 35J] 1 shows the classification of cancer and healthy samples by PCA projection based on metabolite data.

[0089] [Figure 36]The protein, lipid, and metabolite features included in the classifier are shown.

[0090] [Figure 37] The performance of the classifier in a multi-omics study is shown, including the receiver operating characteristic (ROC) curve for pathology classification. Area under the curve (AUC) values ​​are also included in the figure, with 90% confidence intervals shown in parentheses.

[0091] [Figure 38A] The performance of classifiers trained using data from genomics assays is shown, including ROC curves for disease state classification. AUC values ​​at the bottom of the figure are shown ± values ​​based on 90% confidence.

[0092] [Figure 38B] The performance of a classifier trained on data from a genomics assay ("Genomics"), a classifier trained on data from a mass spectrometry assay ("Mass-spec"), and a classifier trained on data from both genomics and mass spectrometry assays ("Combined") is shown. The data shown includes ROC curves for disease state classification. AUC values ​​include ± values ​​based on 90% confidence.

[0093] [Figure 39A] 1 shows a graphical summary of 18 samples from liver cancer subjects used in Example 17.

[0094] [Figure 39B] Coefficient of variation (CV) values ​​for some peptides and proteins obtained in the studies described herein are shown.

[0095] [Figure 39C] 1 shows exemplary protein abundance heat maps of samples from subjects with liver cancer and healthy subjects.

[0096] [Figure 39D]1 shows examples of differences in protein abundance identified in samples from subjects with liver cancer or healthy subjects after contacting the samples with various particles described herein.

[0097] [Figure 39E] 1 includes graphs showing that the lipidomic data obtained from the samples is highly reproducible.

[0098] [Figure 39F] 1 shows that samples from subjects with liver cancer exhibited different lipid profiles than healthy controls. The top 50 lipids based on the p-value of this analysis are displayed for each patient sample.

[0099] [Figure 39G] 1 shows univariate lipid differences in samples from subjects with liver cancer compared to healthy subjects.

[0100] [Figure 40A] 1 shows a graphical summary of nine samples from ovarian cancer subjects used in Example 19.

[0101] [Figure 40B] 1 shows exemplary protein abundance heat maps of samples from subjects with ovarian cancer and healthy subjects.

[0102] [Figure 40C] 1 shows univariate lipid differences in samples from subjects with ovarian cancer compared to healthy subjects.

[0103] [Figure 41] 1 shows examples of steps in screening and treating patients with colon cancer.

[0104] [Figure 42] 1 shows the age and sex breakdown of 268 subjects in the NSCLC biomarker discovery study.

[0105] [Figure 43] Protein counts are shown for each study group, including healthy, comorbid, NSCLC stage 1 "NSCLC_1", NSCLC stage 2 "NSCLC_2", NSCLC stage 3 "NSCLC_3", and NSCLC stage 4 "NSCLC_4".

[0106] [Figure 44] Protein counts for depleted plasma DP and particle panels are shown.

[0107] [Figure 45] 1 shows a summary of differential detection of proteins across subjects versus the mean abundance of said proteins for the particle panel and 10 particle types in depleted plasma (DP).

[0108] [Figure 46] 1 shows the performance of a cross-validated particle panel classifier, where the x-axis shows the proportion of classifications that are false positives and the y-axis shows the proportion of classifications that are true positives.

[0109] [Figure 47] Graphs of random forest models for healthy vs. NSCLC (stages 1, 2, and 3) for depleted plasma (left) and 10 particle panel (right) are shown, with the false positive rate on the x-axis and the true positive rate on the y-axis.

[0110] [Figure 48] Shows the performance of classifier features across study samples.

[0111] [Figure 49] The results are shown for 10 rounds of 10-fold cross-validation with randomized subject class assignment, repeated 10 times, with the false positive rate on the x-axis and the true positive rate on the y-axis.

[0112] [Figure 50] ROC plots of 13 peptides by MRM-MS and 2 proteins by ELISA are shown after removal of proteins found in depleted plasma.

[0113] [Figure 51] Random forest models for all study group comparisons are shown.

[0114] [Figure 52] Some discriminant features for the study group comparison are shown.

[0115] [Figure 53] Protein counts (e.g., number of proteins identified from corona analysis) are shown for panel sizes ranging from 1 particle type to 12 particle types.

[0116] [Figure 54] Examples of biomarkers are shown below.

[0117] [Figure 55] FIG. 1 illustrates a non-limiting example of a web / mobile application delivery system (in this case, a system that delivers a browser-based and / or native mobile user interface); and

[0118] [Figure 56] FIG. 1 illustrates a non-limiting example of a cloud-based web / mobile application delivery system, where the system includes elastically load-balanced, automatically scaled web server and application server resources and synchronized replicated databases.

[0119] [Figure 57] 1 shows the ROC curve of the lung nodule classifier, with the sensitivity and corresponding specificity listed.

[0120] [Figure 58] The feature information and importance of the pulmonary nodule classifier shown in Figure 57 are shown.

[0121] [Figure 59]1 shows some aspects of the samples used in the studies described herein.

[0122]

[0123] [Figure 60] The number of proteins observed in the process control samples is shown.

[0124] [Figure 61] Some coefficient of variation (CV) values ​​are shown.

[0125] [Figure 62] 1 includes protein abundance heat maps of samples from subjects with malignant and benign lung nodules.

[0126] [Figure 63] Includes a volcano diagram plotting the log fold change in protein abundance against the negative log of the p-value.

[0127] [Figure 64] Examples of some proteins from the initial univariate analysis are shown.

[0128] [Figure 65A] 1 includes graphs showing several proteins that were upregulated in biofluid samples from subjects with malignant pulmonary nodules.

[0129] [Figure 65B] 1 includes a graph showing several proteins that were downregulated in biofluid samples from subjects with malignant lung nodules.

[0130] [Figure 66] 1 includes graphs showing that differentially expressed proteins were enriched in metabolic and phosphorylation pathways.

[0131] [Figure 67]Some extrapolated mRNA data showing differentially expressed proteins in metabolic pathways are shown.

[0132] [Figure 68] This is an image showing where some samples were collected for the study.

[0133] [Figure 69A] 1 illustrates some aspects of research subjects and proteomics platforms that can be used in the methods described herein.

[0134] [Figure 69B] 1 illustrates several embodiments of proteomics platforms that can be used in the methods described herein.

[0135] [Figure 69C] Some additional multi-omics aspects are shown.

[0136] [Figure 70] Included is a graphical depiction of the coefficient of variation (CV) values ​​obtained in the studies described herein.

[0137] [Figure 71] Included are empirical power curves for protein changes in the studies described herein.

[0138] [Figure 72] Included is a graphical representation of the detected protein groups and peptide numbers obtained in the studies described herein.

[0139] [Figure 73] 1 includes a graphical representation of protein concentration versus natural log protein intensity data obtained in the studies described herein.

[0140] [Figure 74] 1 includes a graphical representation of protein concentrations for data obtained in the studies described herein.

[0141] [Figure 75A] Contains median normalized log intensity CVs for proteins detected in 100% of samples.

[0142] [Figure 75B] Include the median normalized log intensity CV of proteins detected in at least 25% of the samples.

[0143] [Figure 76] Contains the number of unique protein groups within some sample data.

[0144] [Figure 77A] Relative fluorescence units versus concentration for several standard curves are included.

[0145] [Figure 77B] Includes relative fluorescence units for several standard curves.

[0146] [Figure 78A] Included are peptide yields for some of the nanoparticles used in the experiments described herein.

[0147] [Figure 78B] Included are peptide yields for some of the nanoparticles used in the experiments described herein.

[0148] [Figure 79A] Includes a graph of MS1 intensity over time.

[0149] [Figure 79B] Includes intraday CV of MS1 intensity.

[0150] [Figure 80A] Contains a graph of iRT peptides ranked by FWHM.

[0151] [Figure 80B]A plot showing retention times is included.

[0152] [Figure 81A] A plot showing the distribution of protein group numbers per sample is included.

[0153] [Figure 81B] Includes intraday CV of MS1 intensity.

[0154] [Figure 82] Volcano plots of the intensity differences and P values ​​of peptides detected in biofluid samples are included, with the x-axis displaying the median difference in peptide-level intensities and the y-axis displaying the harmonic mean-based peptide P values.

[0155] [Figure 83] 1 includes graphs showing several transitions for peptide ANVFVQLPR from protein P35858 in benign and malignant groups.

[0156] [Figure 84] Graphs showing the comparison of OpenTarget (OT) scores for lung cancer and significant differences between peptides are included. The graphs show OpenTarget scores on the x-axis and P values ​​on the y-axis.

[0157] [Figure 85] Volcano plots of intensity differences and P-values ​​for metabolites in lung nodule subjects are included. The volcano plots display the median difference in intensity on the x-axis and P-values ​​on the y-axis.

[0158] [Figure 86] Included is a diagram showing the lung discovery sample cohort by seer. The diagram shows that of 589 eligible subjects, 186 subjects met all criteria.

[0159] [Figure 87]A diagram showing the step-by-step approach to discovering a version 1 classifier, a version 2 classifier, and a version 3 classifier through test development is shown.

[0160] [Figure 88] Graphs showing power curves for analyte classes are included. The graphs include curves for proteins, metabolites, and lipids.

[0161] [Figure 89] Volcano plots of intensity differences and P-values ​​for peptides in lung nodule subjects are included. The volcano plots display the median difference in peptide-level intensities on the x-axis and harmonic mean-based peptide p-values ​​on the y-axis.

[0162] [Figure 90] 1 includes a graph showing several transitions for the peptide LEYLLLSR from protein P35858 in benign and malignant groups.

[0163] [Figure 91] 1 includes graphs showing several transitions for peptide ANVFVQLPR from protein P35858 in benign and malignant groups.

[0164] [Figure 92] 1 includes graphs showing several transitions for peptide FLNVLSPR from protein P17936 in benign and malignant groups.

[0165] [Figure 93] An image representing StringDB is shown, highlighting the known interactions of IGFALS and IGFBP3.

[0166] [Figure 94] Volcano plots of intensity differences and P-values ​​for metabolites in lung nodule subjects are included. The volcano plots display the median difference in intensity on the x-axis and P-values ​​on the y-axis.

[0167] [Figure 95] 1 includes a graph showing the amount of biopterin metabolites in benign and malignant groups. The graph displays the type of study group on the x-axis and the amount of metabolite on the y-axis.

[0168] [Figure 96] Volcano plots of intensity difference and P-values ​​for lipids in lung nodule subjects are included. The volcano plots display the median difference in intensity on the x-axis and P-values ​​on the y-axis.

[0169] [Figure 97] Graphs showing the comparison of OpenTarget (OT) scores for lung cancer and significant differences between peptides are included. The graphs show OpenTarget scores on the x-axis and P values ​​on the y-axis.

[0170] [Figure 98] A diagram showing the step-by-step approach to discovering a version 1 classifier, a version 2 classifier, and a version 3 classifier through test development is shown.

[0171] [Figure 99] Included are graphs of the pre-test probability of subjects having benign nodules, and the pre-test and post-test probabilities of subjects having benign nodules. The graphs display probability on the x-axis and number of subjects on the y-axis.

[0172] [Figure 100] Included is a graph comparing sensitivity and specificity, with specificity on the x-axis and sensitivity on the y-axis.

[0173] [Figure 101] 1 shows the ROC curve for 223 subjects with mRNA data in a colorectal cancer (CRC) study. The false positive rate is displayed on the x-axis and the true positive rate is displayed on the y-axis. AUC values ​​are provided.

[0174] [Figure 102]Included are volcano plots showing differential expression of various genes in colorectal cancer studies.

[0175] [Figure 103] The ROC curves for ProteoGraph, mRNA, and ProteoGraph+mRNA are shown, and the respective AUC values ​​are provided.

[0176] [Figure 104] Figure 1 shows the ROC curves for ProteoGraph, PiQuant, mRNA, microRNA, and ProteoGraph+PiQuant+mRNA+microRNA. The respective AUC values ​​are provided.

[0177] [Figure 105] The ROC curves for PiQuant, mRNA, and PiQuant+mRNA are shown, and the respective AUC values ​​are provided.

[0178] [Figure 106] 1 shows ROC curves for classification based on distinct or combined types of biomolecules.

[0179] Incorporation by Reference All publications, patents, and patent applications mentioned herein are incorporated by reference to the same extent as if each individual publication, patent, or patent application was specifically and individually indicated to be incorporated by reference herein. To the extent that the publications and patents or patent applications incorporated by reference conflict with the disclosure contained herein, the present specification is intended to supersede and / or take precedence over any such conflicting material. DETAILED DESCRIPTION OF THE INVENTION

[0180] The present disclosure provides a non-invasive method for diagnosing or ruling out the presence of a disease in a subject or the risk of developing a disease in a subject. The disease may include cancer, such as pancreatic cancer, breast cancer, liver cancer, ovarian cancer, or colon cancer. If early-stage disease in a subject is identified and treatment is provided early, the subject can be saved from further disease progression. Non-invasive tests can also be used to rule out the presence of a disease, thereby avoiding the need for the subject to undergo invasive tests such as biopsies, which may be painful and stressful or may pose a risk of injury to the subject.

[0181] Multi-omics approaches can unlock the ability to detect disease earlier in its onset and improve the accuracy of disease detection. Figure 1A illustrates several embodiments of multi-omics approaches to early disease detection that can combine genomic DNA or DNA methylation information (examples of what may be generally static indicators of risk) with molecular phenotypic information obtained from proteomics or metabolomics, which may be more dynamic indicators of function. Figure 24 also illustrates several embodiments that may be included in multi-omics methods, including several examples of pathologies that can be detected or assessed. Figure 1B illustrates an example of the integration of multiple omics data types. Any of the embodiments in these figures can be used in the methods described herein.

[0182] FIG. 2A shows a non-limiting example of a method for predicting whether a subject has or is at risk for developing a disease, such as cancer. The analysis may include obtaining a biological fluid sample from the subject (200). The sample can be assayed or analyzed. The biological fluid sample can be any one or any combination of the biological fluids described herein. The sample can be either analyzed directly to generate data, such as proteomic data (202); or, prior to analysis in 202, contacted with particles described herein to obtain adsorbed biomolecules (203). After obtaining data from analysis in 202, additional analysis (203) can be performed on the sample obtained from 200 or 201 to obtain additional data sets, such as transcriptomic data, genomic data, metabolomic data, or a combination thereof. The data or data sets obtained from analysis in 202 or 203 can be used to generate a classifier (205). The classifier can be applied to identify whether the subject is likely to have the disease or is at risk for having the disease. The generation or application of the classifier can be further repeated or refined to improve the analysis. Figure 2B further shows some details that can be used in the methods described herein. Either the embodiment of Figure 2A or Figure 2B can be used in the methods described herein, such as the classification method.

[0183] Furthermore, the analysis shown in Figure 2A or Figure 2B can be applied before or during any of the steps included in Figure 3A. For example, the evaluation or analysis can be completed early in the patient's course, before, immediately after, or as part of an invasive workup. It is useful to screen high-risk patients before performing an invasive procedure, such as a biopsy or invasive treatment. In general, opportunities exist where the methods described herein may be useful, such as screening high-risk patients for early detection of disease. The methods described herein can be used for such detection with greater accuracy and convenience than other methods. In Figure 3A, the non-invasive workup may include medical imaging, and the invasive workup may include obtaining a biopsy. The biopsy may be suspicious of a tumor. Similar patient courses for pancreatic cancer, liver cancer, and colon cancer are shown in Figures 3B, 3C, and 41. The evaluation or analysis can be completed at or before any point in Figure 3B, 3C, or 41.

[0184] When the disease is pancreatic cancer, opportunities exist in screening high-risk patients before biopsy or pancreatoscopy. For example, a primary opportunity for using the methods described herein includes screening high-risk pancreatic cancer patients for early detection with improved accuracy and convenience. In the course of liver cancer patients, opportunities exist in screening high-risk liver cancer patients before biopsy. For example, a primary opportunity for using the methods described herein may include improving decisions about indeterminate liver nodules to determine whether a biopsy is necessary. Another opportunity may include surveillance or diagnosis of small, low-risk nodules, or follow-up (e.g., 3-6 months) to track the progression of small nodules. In the course of colorectal cancer (CRC) patients, opportunities exist in screening high-risk patients before colonoscopy. Another opportunity may exist in improving decisions about imaging or biopsy procedures.

[0185] Non-invasively acquired samples can be used to diagnose disease by generating omics data and identifying patterns in the omics data that are associated with the disease. Disease diagnosis can be improved by combining and analyzing multiple types of data (e.g., multiple datasets, such as omics datasets). For example, combining multiple data types can improve the accuracy of predicting whether a subject has a particular disease. When individual datasets independently exhibit errors or do not completely overlap, the combined data may be more accurate than the individual datasets. Methods described herein include generating or obtaining multi-omics data and using the multi-omics data to make predictions about whether a subject has a disease. Various methods for combining or analyzing multi-omics data are described. The use of multi-omics data and disease assessment are described in further detail.

[0186] Several methods can be used to classify lung nodules. Pulmonary nodules can be either benign or malignant. Malignant lung nodules can rapidly progress to lung cancer, a common and fatal cancer. Improved differentiation of malignant and benign lung nodules is needed. On the one hand, early diagnosis of malignant lung nodules can lead to earlier treatment regimens for subjects with malignant lung nodules and improve their prognosis. On the other hand, noninvasive diagnosis of benign or nonmalignant lung nodules can help avoid obtaining lung biopsies, which can be expensive and invasive, and therefore may be better for subjects with nonmalignant lung nodules.

[0187] However, little progress has been made in developing useful clinical tests for diagnosing and analyzing whether lung nodules are benign or malignant. Imaging methods often result in high rates of misdiagnosis (e.g., false positives). Smaller nodules are typically not detected by these imaging methods. Other non-invasive methods, such as biomarker screening, also have limitations. Given that plasma is in contact with many tissues in the body, proteins in plasma can be a useful biomarker discovery matrix. However, plasma proteins can pose challenges due to several factors, including a wide range of concentrations (e.g., 10 orders of magnitude). Complex biochemistry workflows have attempted to circumvent these challenges, but they may not be practical for discovery studies large enough to warrant validation and replication. Alternatively, biomarker research has been limited to the evaluation or reevaluation of existing markers without substantial improvement in clinical performance. Therefore, there remains a need for methods for diagnosing or screening for the presence of benign or malignant lung nodules based on the analysis of biomarkers in biofluid samples. The methods described herein can address this need.

[0188] Disclosed herein are methods that include obtaining biomolecular data. The biomolecular data may include multi-omics data. The method may include generating or receiving the data and then evaluating it using a classifier. The evaluation may include applying the classifier to identify a disease, rule out the presence of a disease, predict the likelihood of a disease, or select a treatment for a disease.

[0189] disease The methods described herein can be used to evaluate a pathological condition. The methods described herein can be used to predict or identify a pathological condition. A pathological condition may include a disease or disorder such as cancer. Examples of cancer include lung cancer, colon cancer, pancreatic cancer, liver cancer, ovarian cancer, breast cancer, prostate cancer, melanoma, bladder cancer, lymphoma, leukemia, kidney cancer, or uterine cancer. In some embodiments, the cancer is breast cancer. A disease may include a disorder. A pathological condition may include having a comorbidity associated with a disease or disorder. A reference to whether a subject has a pathological condition may include the subject being healthy. A healthy state may exclude a pathological condition. For example, a healthy state may exclude having cancer. A pathological condition may exclude being healthy.

[0190] These methods may be useful for cancer diagnosis. These methods may be useful for cancer screening. The method may be useful for cancer treatment. The method may include assaying proteins in a biological fluid sample obtained from a subject having or suspected of having a nodule, such as a lung nodule, to obtain a protein measurement. The method may include applying a classifier to the protein measurement, thereby identifying the protein measurement as indicative of the lung nodule being cancerous or non-cancerous. In some cases, the classifier is generated using proteomic data obtained by contacting a training sample with particles such that the particles adsorb proteins in the training sample and assaying the proteins adsorbed to the particles. Some embodiments include obtaining or receiving a biological fluid sample from a subject.

[0191] In some embodiments, the cancer detected by the methods described herein may be pancreatic cancer, liver cancer, ovarian cancer, or colon cancer. Cancer diagnosis may be improved by obtaining proteomics data or other omics data (such as lipidomics data). Cancer diagnosis can be improved by combining and analyzing multiple types of data (e.g., multiple data sets). For example, combining multiple data types, including proteomics, transcriptomics, genomics, metabolomics, or a combination thereof, can improve the accuracy of predicting whether a subject has cancer. In some embodiments, the methods described herein include generating or obtaining data and using the data to predict whether a subject has cancer. The method may also include identifying the type of cancer (e.g., liver cancer vs. ovarian cancer). Various methods for combining or analyzing data are described, and the use of the data for cancer assessment is further described.

[0192] Cancer can be early stage or late stage. An example of early stage cancer is stage I. Early stage may include stage I or II. Early stage may include stage I, II, or III. An example of late stage cancer may include stage 4.

[0193] The cancer may include pancreatic cancer. The pancreatic cancer may be early-stage pancreatic cancer. In other embodiments, the pancreatic cancer may be late-stage pancreatic cancer. Non-invasively obtained samples can be used to diagnose cancer by generating data and identifying patterns within the data related to cancer, such as pancreatic cancer. In certain embodiments, the method of detecting cancer may include additional screening or diagnostic methods, such as a computed tomography (CT) scan indicative of pancreatic cancer, a magnetic resonance imaging (MRI) scan indicative of pancreatic cancer, a positron emission tomography (PET) scan indicative of pancreatic cancer, an ultrasound indicative of pancreatic cancer, an endoscopic retrograde cholangiopancreatography indicative of pancreatic cancer, an angiogram indicative of pancreatic cancer, a liver function test (LFT) indicative of pancreatic cancer, an elevated carcinoembryonic antigen (CEA) level relative to a control or baseline measurement, an elevated carbohydrate antigen (CA) 19-9 level relative to a control or baseline measurement, or a combination thereof. In some embodiments, methods for detecting pancreatic cancer may include identifying a subject with symptoms such as jaundice, abdominal pain, enlarged gallbladder or liver, blood clots, digestive disorders, or depression, or a combination thereof, any of which may be used in identifying subjects at risk for having pancreatic cancer.

[0194] Cancer may include liver cancer. In some embodiments, the cancer detected by the method described herein may be liver cancer. Liver cancer may be early stage liver cancer. In other embodiments, liver cancer may be late stage liver cancer. In some cases, liver cancer may be stage I, II, III or IV liver cancer. In some cases, the stage of liver cancer is unknown. Non-invasively obtained samples can be used to diagnose cancer by generating data and identifying patterns in the data related to cancer such as liver cancer. In certain embodiments, the method for detecting cancer may include additional screening or diagnostic methods, such as undergoing a dynamic contrast computed tomography (CT) scan indicating liver cancer, a magnetic resonance imaging (MRI) scan indicating liver cancer, undergoing liver function tests (LFTs) indicating liver cancer, elevated bilirubin levels relative to a control or baseline measurement, elevated aminotransferase levels relative to a control or baseline measurement, elevated alkaline phosphatase levels relative to a control or baseline measurement, having hypoalbuminemia, elevated prothrombin time relative to a control or baseline measurement, elevated alpha-fetoprotein levels relative to a control or baseline measurement, or having liver nodules, or a combination thereof. In some embodiments, the method for detecting cancer may include identifying a subject's symptoms, such as abdominal discomfort, pain, and tenderness, jaundice, white, calcareous stools, nausea, vomiting, bruising, or easy bleeding, weakness, or fatigue, or a combination thereof. Any of these embodiments can be used to identify subjects at risk for liver cancer.

[0195] The cancer may include ovarian cancer. In some embodiments, the cancer detected by the methods described herein may be ovarian cancer. The ovarian cancer may be early-stage ovarian cancer. In other embodiments, the ovarian cancer may be late-stage ovarian cancer. In some cases, the stage of ovarian cancer may be unknown. In some embodiments, the stage of ovarian cancer may be stage I, II, III, or IV. Non-invasively obtained samples generate data that can be used to diagnose cancer by identifying patterns in the data related to cancer, such as ovarian cancer. In certain embodiments, the method of detecting cancer may include additional screening or diagnostic methods, such as a computed tomography (CT) scan indicating ovarian cancer, undergoing a magnetic resonance imaging (MRI) scan indicating ovarian cancer, undergoing a positron emission tomography (PET) scan indicating ovarian cancer, undergoing a transvaginal ultrasound indicating ovarian cancer, having elevated cancer antigen (CA)-125 levels relative to a control or baseline measurement, or having an ovarian cyst, or a combination thereof. In some embodiments, the method of detecting cancer may include identifying a subject experiencing symptoms such as pelvic pressure, lower abdominal pain, vaginal bleeding, weight gain, weight loss, irregular menstrual cycles, unexplained lower back pain that worsens over time, increased urination, gas, nausea, vomiting, or loss of appetite, or a combination thereof. Any of these embodiments can be used in identifying subjects at risk for having ovarian cancer.

[0196] The cancer may include colon cancer or colorectal cancer (CRC). In some embodiments, the cancer detected by the methods described herein may be colon cancer. The colon cancer may be early-stage colon cancer. In other embodiments, the colon cancer may be late-stage colon cancer. Non-invasively obtained samples can be used to diagnose cancer by generating data and identifying patterns within the data related to cancer, such as colon cancer. Cancer diagnosis can be improved by obtaining proteomic data. In certain embodiments, the method of detecting cancer may include additional screening or diagnostic methods, such as a computed tomography (CT) scan indicative of colon cancer, liver function tests (LFTs) indicative of colon cancer, measuring carcinoembryonic antigen (CEA) levels relative to a control or baseline measurement, measuring blood in the stool, performing a fecal immunochemical test (FIT), or a combination thereof. Any of these embodiments can be used in identifying subjects at risk for having colon cancer. For example, a subject identified as being at risk for having colon cancer can be identified as being at risk by one of these methods. By using the non-invasive method described herein, patients who do not have colorectal cancer can avoid undergoing further invasive examination or treatment, such as colonoscopy or cancer biopsy, or colorectal cancer treatment.On the other hand, the non-invasive method described herein can be used to identify people who are likely to have colorectal cancer, and confirm that the patient needs to undergo further examination (for example, invasive examination) or treatment.Colorectal cancer can be an example of colorectal cancer (CRC).References or teachings in this specification related to colorectal cancer can be applied to CRC, or vice versa.

[0197] The cancer may include lung cancer. An example of lung cancer is non-small cell lung cancer (NSCLC). An example of lung cancer is small cell lung cancer. A method for diagnosing lung nodules is disclosed. The method may be useful for diagnosing, treating, or screening patients in whom lung nodules have been identified by computed tomography (CT) scan and who have not undergone a lung biopsy. The method may be useful for informing a physician about the likelihood that a lung nodule is benign or malignant. Test results from such a method allow a physician to avoid a patient undergoing an unnecessary biopsy. For example, the method can be used as a rule-out test. Test results from such a method allow a physician to identify subjects that should undergo a biopsy. For example, the method can be used as a rule-in test.

[0198] A diagnostic method for identifying candidates for CT imaging is disclosed. The method may be useful for diagnosing, treating, or screening patients who may be candidates for CT imaging. The method may be useful for high-risk patients (e.g., as defined by the USPSTF or other organizations) who are candidates for lung cancer screening but have not yet undergone a CT scan. The method can inform physicians of the likelihood that the patient has lung cancer. Thus, the method can inform physicians of the patient's pulmonary urgency or the need to obtain that CT scan. Such a method may be useful for high-risk patients, such as those who are not amenable to other CT screening methods. The method can improve patient selection or compliance for CT imaging. The method can improve patient selection or compliance for biopsy.

[0199] A method for recurrent monitoring is disclosed. The method may be useful for monitoring patients with potentially resectable lung cancer. The method may be useful for monitoring patients undergoing post-operative therapeutic intervention. The method may be useful for monitoring patients undergoing adjuvant chemotherapy or radiotherapy intervention. The method may be useful for detecting cancer recurrence before CT scan or other medical imaging. The method may be useful for recurrence surveillance testing. The method may be tailored or developed to suit the patient's treatment regimen.

[0200] Described herein is a method that includes assaying proteins in a biological fluid sample obtained from a subject having or suspected of having a lung nodule to obtain protein measurements; and applying a classifier to the protein measurements to thereby identify the protein measurements as indicative of the lung nodule being cancerous or non-cancerous, where the classifier is generated using proteomic data obtained by contacting a training sample with particles such that the particles adsorb proteins in the training sample and assaying the proteins adsorbed to the particles. This method may be useful for diagnosing or screening for cancer.

[0201] Described herein is a method that includes obtaining a biological fluid sample from a subject having a lung nodule; contacting the biological fluid sample with particles such that the particles adsorb biomolecules, including proteins, to the particles; assaying the biomolecules adsorbed to the particles to generate proteomic data; and classifying the proteomic data as indicative of whether the lung nodule is cancerous or non-cancerous. This method may be useful for diagnosing or screening for cancer.

[0202] Described herein are methods for determining a pulmonary nodule-associated condition in a sample obtained from a subject. In some embodiments, the pulmonary nodule-associated condition includes the presence or absence of a pulmonary nodule in the subject. In some embodiments, the pulmonary nodule-associated condition includes determining whether the pulmonary nodule is benign or malignant. In some embodiments, the method includes screening for a pulmonary nodule-associated condition by assaying a biomarker in a sample obtained from the subject. In some embodiments, the biomarker includes at least one protein in the sample. In some embodiments, the sample is a biological fluid sample. In some embodiments, the biological fluid sample is contacted with particles described herein to adsorb proteins in the biological fluid sample. In some embodiments, the method includes obtaining protein measurements of proteins in the sample. In some embodiments, the method includes applying a classifier to the protein measurements, thereby identifying the protein measurements as indicative of whether the pulmonary nodule is cancerous or non-cancerous. In some embodiments, the classifier is generated using proteomic data obtained by contacting a training sample with particles such that the particles adsorb proteins in the training sample. The adsorbed proteins can then be assayed by methods described herein. In some embodiments, the subject is suspected of having a lung nodule or is identified as having a lung nodule by the imaging methods described herein. In some embodiments, a report is generated based on identifying protein measurements that indicate whether the lung nodule is cancerous or non-cancerous. In some embodiments, the report indicates a likelihood or indication that the lung nodule is cancerous or non-cancerous. In some embodiments, the report indicates that the lung nodule is cancerous. In some embodiments, the report indicates that the lung nodule comprises non-small cell lung cancer (NSCLC). In some embodiments, the methods described herein generate a classifier that includes features that indicate protein measurements that indicate whether the lung nodule is cancerous or non-cancerous.In some embodiments, the features comprise control protein measurements, mass spectra, m / z ratios, chromatographic results, immunoassay results, or light or fluorescence intensities. In some embodiments, the classifier is trained using any one of the computational or machine learning methods described herein.

[0203] Described herein, in some embodiments, is a method for recommending lung cancer treatment to a subject when the subject is determined to have a malignant lung nodule based on an analysis of the protein measurements described herein, in some embodiments, the protein measurements are classified as indicating that the lung nodule is cancerous.

[0204] Disclosed herein in some embodiments are methods useful for diagnosing, screening, or treating a subject. Some embodiments include assaying proteins in a biological fluid sample obtained from a subject suspected of having a lung nodule to obtain protein measurements. Some embodiments include applying a classifier to the protein measurements. Some embodiments include identifying the protein measurements as indicative of the subject having a lung nodule. In some embodiments, the classifier is generated using proteomic data obtained by contacting a training sample with particles such that the particles adsorb proteins in the training sample and assaying the proteins adsorbed to the particles.

[0205] Disclosed herein in some embodiments are methods useful for diagnosing, screening, or treating a subject. Some embodiments include assaying proteins in a biological fluid sample obtained from a subject suspected of having lung cancer to obtain protein measurements. Some embodiments include applying a classifier to the protein measurements. Some embodiments include identifying the protein measurements as indicative of the subject having lung cancer. In some embodiments, the classifier is generated using proteomic data obtained by contacting a training sample with particles such that the particles adsorb proteins in the training sample and assaying the proteins adsorbed to the particles.

[0206] Disclosed herein in some embodiments are methods useful for diagnosing, screening, or treating a subject. Some embodiments include obtaining a biological fluid sample from a subject suspected of having a pulmonary nodule. Some embodiments include contacting the biological fluid sample with particles such that the particles adsorb biomolecules, including proteins, to the particles. Some embodiments include assaying the biomolecules adsorbed to the particles to generate proteomic data. Some embodiments include classifying the proteomic data based on the proteomic data as indicating that the subject has a pulmonary nodule or as not indicating that the subject has a pulmonary nodule.

[0207] Disclosed herein in some embodiments are methods useful for diagnosing, screening, or treating a subject. Some embodiments include obtaining a biological fluid sample from a subject suspected of having lung cancer. Some embodiments include contacting the biological fluid sample with particles such that the particles adsorb biomolecules, including proteins, to the particles. Some embodiments include assaying the biomolecules adsorbed to the particles to generate proteomic data. Some embodiments include classifying the proteomic data based on the proteomic data as indicating that the subject has lung cancer or as not indicating that the subject has lung cancer.

[0208] In some embodiments, a method useful for monitoring a subject is disclosed herein. Some embodiments include obtaining a biological fluid sample from a subject at risk of lung cancer recurrence. Some embodiments include contacting the biological fluid sample with particles such that the particles adsorb biomolecules, including proteins, to the particles. Some embodiments include assaying the biomolecules adsorbed to the particles to generate proteomic data. Some embodiments include classifying the proteomic data based on the proteomic data as indicating that the subject has lung cancer recurrence or as not indicating that the subject has lung cancer recurrence. In some embodiments, the subject is undergoing lung cancer treatment, such as chemotherapy, radiation therapy, or surgery. In some embodiments, the cancer may be resectable. In some embodiments, the lung cancer includes NSCLC.

[0209] In some cases, lung nodules are described as malignant or cancerous. The terms malignant and cancerous may be used interchangeably. A malignant or cancerous lung nodule may be called a lung cancer, and vice versa. In some cases, lung nodules are described as benign or non-cancerous. The terms benign and non-cancerous may be used interchangeably.

[0210] Samples and subjects Some embodiments relate to a subject. For example, the method described herein can be used to evaluate a subject or to evaluate a sample from a subject. Multi-omics data can be generated from the subject's sample.

[0211] The methods described herein can be used to identify subjects who are likely to have or are at risk of having a disease such as cancer. The subject may have lung cancer, pancreatic cancer, liver cancer, ovarian cancer, or colon cancer. The cancer may include adenocarcinoma, such as pancreatic adenocarcinoma. The subject may have cancer. The subject may not have cancer. The subject may have pancreatic cancer, liver cancer, ovarian cancer, or colon cancer. The subject may not have pancreatic cancer, liver cancer, ovarian cancer, or colon cancer. The subject may be at risk of having pancreatic cancer, liver cancer, ovarian cancer, or colon cancer. The subject may have a mass (e.g., a nodule or cyst) in the pancreas. The subject may have a mass (e.g., a nodule) in the liver. Liver cancer may include hepatocellular carcinoma (HCC). Liver cancer may include stage I, stage II, stage III, or stage IV liver cancer. The subject may have a mass (e.g., a nodule or cyst) in one or both ovaries. The ovarian cancer may include stage I, stage II, stage III, or stage IV ovarian cancer. The ovarian cancer may include stage III ovarian cancer. The ovarian cancer may include stage IV ovarian cancer. The subject may have a mass (e.g., a nodule) in the colon. The subject may have lung nodule cancer. The subject may be at risk of having breast cancer. The subject may have a mass (e.g., a nodule or cyst) in the breast.

[0212] The sample can be obtained from a subject for the purpose of identifying cancer in the subject. The subject may be suspected of having cancer or may not have cancer. The method can be used to confirm or refute the suspicion of cancer.

[0213] The data described herein can be generated from a subject's sample. The sample can be a biological fluid sample or a mass sample (e.g., an abnormal growth biopsied from a subject). Examples of biological fluids include blood, serum, or plasma. The sample may include a blood sample. The sample may include a serum sample. The sample may include a plasma sample. One or more biological fluid samples may include a blood, serum, or plasma sample. Other examples of biological fluids include urine, tears, semen, milk, vaginal fluid, mucus, saliva, sweat, or a cell homogenate.

[0214] The sample can be obtained from a subject for the purpose of identifying a pathological condition in the subject. The subject may be suspected of having a pathological condition or may not have a pathological condition. This method can be used to confirm or refute the suspicion of a pathological condition. In some embodiments, the sample from the subject is used to determine whether a mass, nodule (e.g., a lung nodule), or cyst is cancerous or non-cancerous.

[0215] A biological fluid sample can be obtained from a subject. For example, a blood, serum, or plasma sample can be obtained from a subject by drawing blood. Other methods for obtaining a biological fluid sample include aspiration or swabbing.

[0216] A biological fluid sample can be acellular or substantially acellular. To obtain a cell-free or substantially acellular biological fluid sample, the biological fluid can be subjected to sample preparation methods such as centrifugation and pellet removal.

[0217] A non-biological fluid sample can be obtained from a subject or patient. For example, the sample may include a tissue sample. Some examples of organs or tissues that can be sampled include lung, colon, pancreas, liver, breast, or ovarian tissue. The sample may include a mass taken from a subject's organ or tissue. The mass may be suspected to be cancerous. The mass may include a nodule (e.g., a colon nodule or liver nodule). The mass may include a cyst (e.g., an ovarian cyst). The nodule or cyst can be identified by a physician as having a high or low risk of being cancerous prior to performing the methods described herein. The mass may be biopsied, for example, by a needle biopsy procedure. A needle biopsy procedure may involve inserting a thin needle through the subject's abdomen into the liver to obtain a tissue sample, which can then be examined under a microscope for signs of cancer. The sample may include a cell sample. The sample may include a cell or tissue homogenate. The sample may include the supernatant of a centrifuged cell or tissue homogenate.

[0218] The sample may include lung tissue. The sample may include colon tissue. The sample may include pancreatic tissue. The sample may include liver tissue. The sample may include breast tissue. The sample may include ovarian tissue. The tissue may be cancerous. The tissue may be non-cancerous. The tissue may be suspected of being cancerous. The tissue may be malignant. The tissue may be non-malignant. The tissue may be suspected of being malignant.

[0219] A sample (e.g., a biological fluid or tissue sample) can be obtained from a subject during any stage of a screening procedure, such as before, during, or after the stages shown in Figure 3A. The sample can be obtained before or during a stage in which the subject is a candidate for biopsy, pancreatoscopy, or colonoscopy for early detection of disease. The sample can be obtained before or during a non-invasive workup, an invasive workup, treatment, or monitoring stage.

[0220] Data can be generated from a single sample or from multiple samples. Data from multiple samples can be obtained from the same subject. In some cases, different data types are obtained from samples collected differently or in separate containers. Samples may be collected in a container containing one or more reagents, such as a preservation reagent or a biomolecule isolation reagent. Some examples of reagents include heparin, ethylenediaminetetraacetic acid (EDTA), citrate, an anti-lysis agent, or a combination of reagents. Samples from a subject may be collected in multiple containers containing different reagents, such as to preserve or isolate different types of biomolecules. Samples may also be collected in containers that do not contain reagents. Samples can be collected at the same time (e.g., the same time or day) or at different times. Samples can be frozen, refrigerated, heated, or stored at room temperature.

[0221] The methods described herein can be used to identify a subject as likely or unlikely to have a pathological condition. The pathological condition can include cancer, such as pancreatic cancer, liver cancer, ovarian cancer, or colon cancer. Some aspects of the present disclosure include identifying a lung nodule in a subject as cancerous or non-cancerous. The lung nodule can be present in the subject's lung. The subject can be identified as having a lung nodule. In some aspects, the subject has multiple lung nodules. The subject can have lung cancer. The subject can be at risk for lung cancer. The subject can have a pulmonary complication. The subject can have a comorbidity as described herein. The subject can experience shortness of breath. The subject can have fluid in the lung.

[0222] In some cases, the subject is monitored. For example, information about the likelihood that the subject has a pathological condition can be used to determine whether the subject is monitored without providing treatment. In other situations, the subject can be monitored while receiving treatment to see if the subject's pathological condition improves. In some embodiments, a subject with a pulmonary nodule can be monitored to determine the progression of the pulmonary nodule. The subject's pulmonary nodule can be monitored. The subject can be treated as described herein.

[0223] The subject may be a vertebrate. The subject may be a mammal. Mammals may include rats, mice, gerbils, guinea pigs, or hamsters. Mammals may include foxes, bears, dogs, monkeys, cows, pigs, or sheep. The subject may be a primate. Primates may include apes or monkeys. Primates may include chimpanzees, lemurs, bonobos, orangutans, or baboons. The subject may be a human. The subject may be an adult (e.g., at least 18 years of age). The subject may be male. The subject may be female. The subject may have a pathological condition. For example, the subject may have a disease or disorder, co-morbidities of a disease or disorder, or may be healthy.

[0224] The methods described herein may include the use of a sample, such as a biological sample. For example, the method may include determining one or more biomarker measurements in the sample. The biological sample may be from a subject, such as a subject with a pulmonary nodule. The biological sample may include a blood sample from which red blood cells have been removed. For example, the biological sample may include a plasma sample. The biological sample may include a serum sample. The biological sample may include blood or blood components. The biological sample may include a blood sample. The sample described or used herein may be from a subject described herein, such as a subject with an identified pulmonary nodule.

[0225] The sample is consistent with the methods disclosed herein for assessing the presence or absence of one or more biomarkers associated with the presence or absence of a pulmonary nodule or a malignant condition. The subject may be a human or non-human animal. The biological sample may be a biological fluid. For example, the biological fluid may be plasma, serum, CSF, urine, tears, cell lysate, tissue lysate, cell homogenate, tissue homogenate, nipple aspirate, fecal sample, synovial fluid, whole blood, or saliva. The sample may also be a non-biological sample, such as water, milk, a solvent, or one homogenized in a fluid state. The biological sample may contain multiple proteins or proteomic data, which can be analyzed after protein adsorption to the surface of various types of particles in a panel and subsequent digestion of the protein corona. Proteomic data may include nucleic acids, peptides, or proteins. Any of the samples herein may contain a number of different analytes that can be analyzed using the methods disclosed herein. Analytes may be proteins, peptides, small molecules, nucleic acids, metabolites, lipids, or any molecule that may bind to or interact with the surface of a particle type.

[0226] The sample may be a biological fluid. The biological sample may include a biological fluid sample, such as cerebrospinal fluid (CSF), synovial fluid (SF), urine, plasma, serum, tears, gingival crevicular fluid, semen, whole blood, milk, nipple aspirate, ductal lavage, vaginal fluid, nasal fluid, ear fluid, gastric fluid, pancreatic fluid, trabecular fluid, pulmonary lavage, prostatic fluid, sputum, feces, bronchial lavage, fluid from swabbing, bronchial aspirate, sweat, or saliva. The biological fluid may also be a mobilized solid, such as a tissue homogenate, or a fluid extracted from a biological sample. The biological sample may be, for example, a tissue sample or a fine needle aspiration (FNA) sample. The biological sample may also be a cell culture sample. For example, a sample that can be used in the methods disclosed herein may contain cells growing in cell culture or acellular material harvested from a cell culture. The biological fluid may also be a mobilized biological sample. For example, the biological fluid may be a fluidized cell culture extract. The sample may be extracted from a fluid sample, or the sample may be extracted from a solid sample. For example, the sample may include gaseous molecules extracted from a fluidized solid (e.g., volatile organic compounds). In some embodiments, the biological fluid includes blood, plasma, or serum.

[0227] Methods consistent with the present disclosure can include collecting (e.g., isolating, enriching, or purifying) species from a biological sample. The species can be a biomolecule (e.g., a protein), a biopolymer structure (e.g., a peptide aggregate or a ribosome), a cell, or a tissue. The species can be selectively collected from the biological sample. For example, the method can include isolating cancer cells from tissue (e.g., as a tissue biopsy) or from a bodily fluid such as whole blood, plasma, or buffy coat (e.g., as a liquid biopsy). The method can include a sample that does not contain cancer cells. The species can be treated prior to analysis. For example, proteins can be reduced and degraded, nucleic acids can be separated from histones, or cells can be lysed.

[0228] Biological sample can be obtained or derived from human subject.Before processing, biological sample can be stored under various storage conditions, such as different temperatures (for example, at room temperature, under refrigerated or frozen conditions, at 25 ℃, 4 ℃, -18 ℃, -20 ℃ or -80 ℃) or different suspensions (for example, EDTA collection tube, cell-free RNA collection tube or cell-free DNA collection tube).

[0229] In some cases, samples may be depleted prior to biomarker analysis. Samples can be depleted using commercially available kits. For example, kits that can be used to deplete samples can be spin-column-based depletion kits, albumin depletion kits, immunodepletion kits, or abundant protein depletion kits. Non-limiting examples of kits that can be used for sample depletion include the PureProteome™ Human Albumin / Immunoglobulin Depletion Kit (EMD Millipore Sigma), ProteoPrep® Immunoaffinity Albumin & IgG Depletion Kit (Millipore Sigma), Seppro® Protein Depletion Kit (Millipore Sigma), Top 12 Abundant Protein Depletion Spin Columns (Pierce), or Proteome Purify™ Immunodepletion Kit (R&D Systems). Depletion can remove highly abundant biomolecules from a sample. For example, the method may include removing albumin from the plasma sample prior to low concentration biomarker analysis. The sample may comprise depleted plasma.

[0230] Data Generation and Use The methods disclosed herein may include obtaining data, such as multi-omics data, generated from one or more biological fluid samples collected from a subject. The data may include biomolecular measurements, such as protein measurements, transcript measurements, genetic material measurements, or metabolite measurements. Omics data may include any of the following: proteomics data, genomics data, transcriptomics data, or metabolomics data. This section includes several methods for generating each of these types of omics data. Methods for generating or analyzing omics data can also be applied to methods for generating or analyzing individual biomolecules or sets of biomolecules. Other types of omics data can also be generated. Descriptions of generating or analyzing omics data can also be applied to methods for generating or analyzing individual biomolecules or sets of biomolecules that do not necessarily include omics data. Aspects described with respect to biomolecular data may also relate to biomolecular measurements, or vice versa. Data can be labeled or identified as indicative of disease or not indicative of disease. The data can be labeled or identified as indicative of pancreatic cancer, liver cancer, ovarian cancer, or colon cancer, or as not indicative of pancreatic cancer, liver cancer, ovarian cancer, or colon cancer. The methods described herein can include obtaining multi-omic measurements, such as by performing assays.

[0231] The methods described herein may include generating or using omics data. Omics data may include data on all biomolecules of a particular type, such as proteins, transcripts, genetic material, or metabolites. Omics data may also include data on a subset of biomolecules. For example, omics data may include data on 500 or more, 750 or more, 1000 or more, 2500 or more, 5000 or more, 10,000 or more, or 25,000 or more biomolecules of a particular type. The methods described herein can include obtaining measurements of more than 10, more than 20, more than 30, more than 40, more than 50, more than 75, more than 100, more than 250, more than 500, more than 750, more than 1000, more than 1250, more than 2500, more than 5000, more than 7500, more than 10,000, more than 12,500, more than 15,000, more than 17,500, more than 20,000, more than 22,500, or more than 25,000 biomolecules of a particular type. The methods described herein can include obtaining measurements of fewer than 10, 20, 30, 40, 50, 75, 100, 250, 500, 750, 1000, 1250, 2500, 5000, 7500, 10,000, 12,500, 15,000, 17,500, 20,000, 22,500, or 25,000 biomolecules of a particular type. Any of the foregoing numbers of biomolecules can be measured for each of the multiple data types. Multi-omics data includes at least 100 measurements of each of at least two types of omics data. Multi-omics data includes at least 500 measurements of each of at least two types of omics data. Multi-omics data includes at least 1,000 measurements of each of at least two types of omics data. The data may relate to the presence, absence, or amount of a given biomolecule. Example data types may include lipid, protein, peptide, transcript, mRNA, miRNA, DNA sequence, methylation, or metabolite data.

[0232] Other documents are individually and separately indicated to be incorporated by reference for all purposes.

[0233] Deep proteome coverage is advantageous for multi-omics approaches. New technologies and sample availability address past challenges in scaling proteomics. Some challenges include access to large, well-collected and annotated sample cohorts for specific clinical questions, technical challenges associated with plasma proteomics that may limit clinical translation, such as reproducibility, throughput, and depth of coverage, and reproducible measurement and integration of multi-omics datasets that provide novel insights into cancer biology.

[0234] The concepts described herein may help address some of these challenges, for example, by using particles or including additional omics types to address these concerns.

[0235] Methods for multi-omic analysis are disclosed herein. "Multi-omic(s)" or "multiomic(s)" may include analytical approaches for large-scale analysis of biomolecules, where the datasets are multiple omes, such as proteomes, genomes, transcriptomes, lipidomes, and metabolomes. Non-limiting examples of multi-omic data include proteomics data, genomics data, lipidomics data, glycomics data, transcriptomics data, or metabolomics data. The "biomolecule" in the "biomolecular corona" can refer to any molecule or biological component that can be produced by or is present in a living tissue. Non-limiting examples of biomolecules include proteins (protein coronas), polypeptides, polysaccharides, sugars, lipids, lipoproteins, metabolites, oligonucleotides, nucleic acids (DNA, RNA, microRNA, plasmids, single-stranded nucleic acids, double-stranded nucleic acids), metabolomes, and small molecules, such as primary metabolites, secondary metabolites, and other natural products, or any combination thereof. In some embodiments, the biomolecule is selected from the group of proteins, nucleic acids, lipids, and metabolites.

[0236] Some aspects that may be included in a multi-omics strategy may include well-defined disease biobanks with multiple sample types optimized for multi-omics measurements, development and optimization of new proteomic technologies to increase proteome coverage and throughput without compromising reproducibility, or unbiased multi-omics platforms incorporating state-of-the-art instrumentation and advanced machine learning analytics to transform the early detection of complex diseases.

[0237] Proteomics Data Data such as the multi-omics data described herein may include protein data or proteomics data. Proteomics data may include data regarding proteins, peptides, or proteoforms. The data may include only peptides or proteins, or a combination of both. An example of a peptide is an amino acid chain. An example of a protein is a peptide or a combination of peptides. For example, a protein can include one, two, or more peptides linked together. A protein may be a secreted protein. Proteomics data may include data regarding various proteoforms. Proteoforms can include various forms of a protein generated from a genome with various sequence variations, splice isoforms, or post-translational modifications. Proteomics data may be generated using an unbiased, untargeted approach or may include a set of specific proteins. Aspects described with respect to proteomics data may also be relevant to protein data, and vice versa.

[0238] Proteomic data may include information regarding the presence, absence, or amount of various proteins or peptides. For example, proteomic data may include the amount of a protein. The amount of a protein may be expressed as the concentration or amount of the protein, for example, the concentration of the protein in a biological fluid. The amount of a protein may be compared to another protein or another biological molecule. Proteomic data may include information regarding the presence of a protein or peptide. Proteomic data may include information regarding the absence of a protein or peptide. Proteomic data may be differentiated by subtypes, each subtype including different types of proteins, peptides, or proteoforms.

[0239] Proteomic data typically includes data on a large number of proteins or peptides. For example, proteomic data may include information regarding the presence, absence, or quantity of 1,000 or more proteins or peptides. In some cases, proteomic data may include information regarding the presence, absence, or quantity of 5,000, 10,000, 20,000, or more peptides, proteins, or proteoforms. Proteomic data can include up to approximately 1 million proteoforms. Proteomic data may include a range of proteins, peptides, or proteoforms, defined by any of the aforementioned numbers of proteins, peptides, or proteoforms. Some examples of proteins or peptides that can be included in proteomic data are shown in Figures 6, 7, 10B, or 15.

[0240] In some embodiments, the multi-omics data comprises measurements of more than 10 peptides or proteins, more than 15 peptides or proteins, more than 20 peptides or proteins, more than 25 peptides or proteins, more than 30 peptides or proteins, more than 35 peptides or proteins, more than 40 peptides or proteins, more than 45 peptides or proteins, more than 50 peptides or proteins, more than 75 peptides or proteins, more than 100 peptides or proteins, more than 250 peptides or proteins, more than 500 peptides or proteins, more than 1,000 peptides or proteins, more than 2,500 peptides or proteins, more than 5,000 peptides or proteins, more than 10,000 peptides or proteins, more than 15,000 peptides or proteins, or more than 20,000 peptides or proteins. In some embodiments, the multi-omics data includes measurements of at least about 10 peptides or proteins, at least about 15 peptides or proteins, at least about 20 peptides or proteins, at least about 25 peptides or proteins, at least about 30 peptides or proteins, at least about 35 peptides or proteins, at least about 40 peptides or proteins, at least about 45 peptides or proteins, at least about 50 peptides or proteins, at least about 75 peptides or proteins, at least about 100 peptides or proteins, at least about 250 peptides or proteins, at least about 500 peptides or proteins, at least about 1,000 peptides or proteins, at least about 2,500 peptides or proteins, at least about 5,000 peptides or proteins, at least about 10,000 peptides or proteins, at least about 15,000 peptides or proteins, or at least about 20,000 peptides or proteins.In some embodiments, the protein data includes measurements of 10 or fewer peptides or proteins, 15 or fewer peptides or proteins, 20 or fewer peptides or proteins, 25 or fewer peptides or proteins, 30 or fewer peptides or proteins, 35 or fewer peptides or proteins, 40 or fewer peptides or proteins, 45 or fewer peptides or proteins, 50 or fewer peptides or proteins, 75 or fewer peptides or proteins, 100 or fewer peptides or proteins, 250 or fewer peptides or proteins, 500 or fewer peptides or proteins, 1,000 or fewer peptides or proteins, 2,500 or fewer peptides or proteins, 5,000 or fewer peptides or proteins, 10,000 or fewer peptides or proteins, 15,000 or fewer peptides or proteins, or 20,000 or fewer peptides or proteins. A peptide or protein group can comprise or consist of a peptide. A peptide or protein group can comprise or consist of a protein group.

[0241] A protein may also include a post-translational modification (PTM). An example of a PTM may include glycosylation. A protein or peptide may include a glycoprotein or glycopeptide. A protein may include a glycoprotein. A peptide may include a glycopeptide. An example of a PTM may include phosphorylation. A protein or peptide may include a phosphoprotein or phosphopeptide. A protein may include a phosphoprotein. A peptide may include a phosphopeptide.

[0242] Proteomic data can be generated by any of a variety of methods. Generating proteomic data can include using a detection reagent that binds to the peptide or protein to generate a detectable signal. After using the detection reagent that binds to the peptide or protein to generate a detectable signal, a readout can be obtained that indicates the presence, absence, or amount of the protein or peptide. Generating proteomic data can include concentrating, filtering, or centrifuging the sample.

[0243] Proteomic data can be generated using mass spectrometry, chromatography, liquid chromatography, high-performance liquid chromatography, solid-phase chromatography, lateral flow assays, immunoassays, enzyme-linked immunosorbent assays, Western blots, dot blots, or immunostaining, or a combination thereof. Some examples of methods for generating proteomic data include the use of mass spectrometry, protein chips, or reverse-phase protein microarrays. Proteomic data can also be generated using immunoassays, such as enzyme-linked immunosorbent assays, Western blots, dot blots, or immunohistochemistry assays. Generating proteomic data can include the use of immunoassay panels.

[0244] One method for obtaining proteomic data involves the use of mass spectrometry. An example of a mass spectrometry method is the use of high-resolution two-dimensional electrophoresis to separate proteins in parallel from different samples, followed by the selection or staining of differentially expressed proteins that are identified by mass spectrometry. Another method uses stable isotope tags to differentially label proteins from two different complex mixtures. Proteins within the complex mixture are isotopically labeled and then digested to yield labeled peptides. The labeled mixtures are then combined, and the peptides can be separated by multidimensional liquid chromatography and analyzed by tandem mass spectrometry. Mass spectrometry methods can include the use of liquid chromatography-mass spectrometry (LC-MS), a technique that combines the physical separation capabilities of liquid chromatography (e.g., HPLC) with mass spectrometry.

[0245] Proteins can be enriched before being assayed or measured. Enrichment can involve enriching one set of proteins and not another set, or enriching a single protein and not another. Enrichment can be achieved through the use of affinity reagents, for example, by incubating the affinity reagent with the sample before measuring the proteins in the sample. The affinity reagent can include antibodies. The affinity reagent can include particles, such as nanoparticles. Proteins can be adsorbed to the affinity reagent, separated from the remainder of the sample, and then assayed using the proteomic assays described herein.

[0246] Generating proteomic data can include contacting a sample with particles such that the particles adsorb biomolecules, including proteins. The adsorbed proteins can be part of a biomolecular corona. The adsorbed proteins can be measured or identified in generating the proteomic data.

[0247] The generation of proteomic data may include the use of a known amount of an internal reference protein. The reference protein may be labeled. The label may include an isotopic label. The generation of proteomic data may include the use of a known amount of an isotopically labeled internal reference protein (referred to as "PiQuant"). The internal reference protein may be spiked into a sample. The internal reference protein can be used to identify the mass spectra of individual endogenous proteins. The internal reference protein can be used as a standard to determine the amount of individual endogenous proteins. Proteomic measurements can be generated based on the amount of protein added to one of one or more biological fluid samples. Proteomic measurements can be generated based on the amount of labeled protein added to one of one or more biological fluid samples.

[0248] Transcriptomic data Data, such as the multi-omics data described herein, can include transcriptional data or transcriptomics data. Transcriptomics data can include data on nucleotide transcripts, such as RNA. Examples of RNA include messenger RNA (mRNA), ribosomal RNA (rRNA), signal recognition particle (SRP) RNA, transfer RNA (tRNA), small nuclear RNA (snRNA), small nucleolar RNA (snoRNA), long non-coding RNA (lncRNA), microRNA (miRNA), non-coding RNA (ncRNA), or piwi-interacting RNA (piRNA), or a combination thereof. RNA can include mRNA. RNA can include miRNA. Transcriptomics data can be distinguished by subtype, with each subtype containing a different type of RNA or transcript. For example, mRNA data can be included in one subtype, and data on one or more types of small non-coding RNA, such as miRNA or piRNA, can be included in another subtype. miRNA can include 5p miRNA or 3p miRNA.

[0249] Transcriptomic data may include information regarding the presence, absence, or amount of various RNAs. For example, transcriptomic data may include the amount of RNA. The amount of RNA may be expressed as a concentration or number of RNA molecules, e.g., the concentration of RNA in a biological fluid. The amount of RNA may be compared to another RNA or another biological molecule. Transcriptomic data may include information regarding the presence of RNA. Transcriptomic data may include information regarding the absence of RNA. An embodiment described with respect to transcriptomic data may also be relevant to transcript or RNA data, and vice versa.

[0250] Transcriptomic data generally includes data on a large number of RNAs. For example, transcriptomic data may include information on the presence, absence, or quantity of 1,000 or more RNAs. In some cases, transcriptomic data may include information on the presence, absence, or quantity of 5,000, 10,000, 20,000, or more RNAs. Transcriptomic data may also include up to approximately 200,000 transcripts. Transcriptomic data may include a set of transcripts defined by any of the aforementioned numbers of RNAs or transcripts. Some examples of mRNAs that may be included in transcriptomic data are shown in Figure 10B or Figure 15. Some examples of microRNAs that may be included in transcriptomic data are shown in Figure 11B or Figure 15.

[0251] Some examples of mRNAs that can be used as biomarkers are shown in Figure 10B. mRNAs 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 included in Figure 10B can be used as biomarkers in determining whether, for example, a lung nodule is cancerous, or determining the likelihood of such. Some examples of microRNAs that can be used as biomarkers are shown in Figure 11B. MicroRNAs 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 included in Figure 11B can be used as biomarkers in determining whether, for example, a lung nodule is cancerous, or determining the likelihood of such.

[0252] Transcriptomic data can be generated by any of a variety of methods. Generating transcriptomic data can include using a detection reagent that binds to RNA and generates a detectable signal. After using a detection reagent that binds to RNA and generates a detectable signal, a readout can be obtained that indicates the presence, absence, or amount of RNA. Generating transcriptomic data can include concentrating, filtering, or centrifuging the sample.

[0253] Transcriptomic data may include RNA sequence data. Some example methods for generating RNA sequence data include the use of sequencing, microarray analysis, hybridization, polymerase chain reaction (PCR), or electrophoresis, or a combination thereof. Microarrays can be used to generate transcriptomic data. PCR can be used to generate transcriptomic data. PCR may include quantitative PCR (qPCR). Such methods can include the use of detectable probes (e.g., fluorescent probes) that intercalate into double-stranded nucleotides or bind to target nucleotide sequences. PCR may include reverse transcriptase quantitative PCR (RT-qPCR). Generating transcriptomic data may include the use of PCR panels.

[0254] RNA sequence data can be generated by sequencing the target RNA or by first converting the target RNA to DNA (e.g., complementary DNA (cDNA)) and then sequencing the DNA. Sequencing can include massively parallel sequencing. Examples of massively parallel sequencing techniques include pyrosequencing, sequencing by reversible terminator chemistry, sequencing by ligation mediated by ligase enzymes, or phospho-linked fluorescent nucleotides or real-time sequencing. Generating transcriptomics data can also include preparing samples or templates for sequencing. Reverse transcriptase can be used to convert RNA to cDNA. Some template preparation methods include the use of amplification templates derived from single RNA or cDNA molecules or single RNA or cDNA molecule templates. Examples of amplification methods include emulsion PCR, rolling circle amplification, or solid-phase amplification.

[0255] In addition to any of the above methods, generating transcriptomic data may include contacting the sample with particles such that the particles adsorb biomolecules, including RNA. The adsorbed RNA may be part of a biomolecular corona. The adsorbed RNA may be measured or identified in generating the transcriptomic data.

[0256] Genomics Data Data such as multi-omics data described herein may include data regarding genetic material or genomics data. Genomics data may include data regarding genetic material such as nucleic acids or histones. Nucleic acids may include DNA. Genomics data may include information regarding the presence, absence, or amount of genetic material. The amount of genetic material may be expressed as a concentration, absolute number, or may be relative. Aspects described with respect to genomics data may also relate to nucleic acid or DNA data, and vice versa. Nucleic acid data may include RNA data, or genomics data may include transcriptomics data.

[0257] Genomic data may include DNA sequence data. Sequence data may include gene sequences. For example, genomics data may include sequence data for up to approximately 20,000 genes. Genomics data may also include sequence data for non-coding DNA regions. DNA sequence data may include information regarding the presence, absence, or quantity of DNA sequences. DNA sequence data may include information regarding the presence or absence of mutations, such as single nucleotide polymorphisms. DNA sequence data may include DNA measurements of the amount of mutant DNA, e.g., measurements of mutant DNA from cancer cells.

[0258] Genomics data may include epigenetic data. Examples of epigenetic data include DNA methylation data, DNA hydroxymethylation data, or histone modification data. Epigenetic data may include DNA methylation or hydroxymethylation. DNA methylation or hydroxymethylation can be measured globally or regionally within DNA. Methylated DNA may include methylated cytosine (e.g., 5-methylcytosine). Cytosine is often methylated at CpG sites and may indicate gene activation.

[0259] The epigenetic data may include histone modification data. The histone modification data may include the presence, absence, or amount of histone modification. Examples of histone modifications include serotonination, methylation, citrullination, acetylation, or phosphorylation. Some specific examples of histone modifications include lysine methylation, glutamine serotonination, arginine methylation, arginine citrullination, lysine acetylation, serine phosphorylation, threonine phosphorylation, or tyrosine phosphorylation. The histone modification may indicate gene activation.

[0260] Genomics data can be distinguished by subtypes, with each subtype containing different types of genomics data. For example, DNA sequence data can be included in another subtype, epigenetic data can be included in one subtype, or different types of epigenetic data can be included in different subtypes.

[0261] Genomics data can be generated by any of a variety of methods. Generating genomics data may include using a detection reagent that binds to genetic material, such as DNA or histones, to generate a detectable signal. After using a detection reagent that binds to the genetic material and generates a detectable signal, a readout can be obtained that indicates the presence, absence, or amount of the genetic material. Generating genomics data may also include concentrating, filtering, or centrifuging the sample.

[0262] Some examples of methods for generating DNA sequence data include sequencing, microarray analysis (e.g., SNP microarray), hybridization, polymerase chain reaction, or electrophoresis, or a combination thereof. DNA sequence data can be generated by sequencing the target DNA. Sequencing can include massively parallel sequencing. Examples of massively parallel sequencing techniques include pyrosequencing, sequencing by reversible terminator chemistry, sequencing by ligation mediated by ligase enzymes, or phospho-linked fluorescent nucleotides or real-time sequencing. Generating genomics data can also include preparing samples or templates for sequencing. Some template preparation methods include the use of amplification templates derived from single DNA molecules or single DNA molecule templates. Examples of amplification methods include emulsion PCR, rolling circle amplification, or solid-phase amplification.

[0263] DNA methylation can be detected using mass spectrometry, methylation-specific PCR, bisulfite sequencing, HpaII tiny fragment enrichment by ligation-mediated PCR assay, GlaI hydrolysis and ligation adapter-dependent PCR assay, chromatin immunoprecipitation (ChIP) assay combined with DNA microarray (ChIP-on-chip assay), restriction enzyme landmark genomic scanning, methylated DNA immunoprecipitation, pyrosequencing of bisulfite-treated DNA, molecular break-write assay of DNA adenine methyltransferase activity, methyl-sensitive Southern blotting, methyl-CpG binding protein, high-resolution melting analysis, methylation-sensitive single-nucleotide primer extension assay, another methylation assay, or a combination thereof.

[0264] Histone modifications can be detected using mass spectrometry or immunoassays, enzyme-linked immunosorbent assays, Western blots, dot blots, or immunostaining, or a combination thereof.

[0265] In addition to any of the above methods, generating genomics data may include contacting the sample with particles such that the particles adsorb biomolecules, including genetic material. The adsorbed genetic material may be part of a biomolecular corona. The adsorbed genetic material may be measured or identified in generating the genomics data.

[0266] Figure 23 provides embodiments that may be related to transcriptomics or genomics data. The data may include circulating free DNA (cfDNA) methylation, mRNA, miRNA, circulating free miRNA (cf-miRNA), or whole-exome sequencing data. Any sample type, isolation method, quality control (QC) aspect, or sequencing depth provided in the figure may be included. Any embodiment shown in the figure may be included in or used to generate data such as multi-omics data.

[0267] Lipidomics Data Data such as the multi-omics data described herein may include lipid data or lipidomics data. Lipidomic data may include information about the presence, absence, or amount of various lipids. For example, lipidomics data may include the amount of lipid. The amount of lipid may be expressed as the concentration or amount of lipid, for example, the concentration of lipid in a biological fluid. The amount of lipid may be compared to another lipid or another biomolecule. Lipidomic data may include information about the presence of lipid. Lipidomic data may include information about the absence of lipid. Lipid or lipidomics data may be included in metabolite or metabolomics data. Aspects described with respect to lipidomics data may also be relevant to lipid data, and vice versa.

[0268] Many organisms contain complex lipids (e.g., humans express over 600 types of lipids), and their relative expression can serve as powerful markers for determining biological and health states. Lipids are a diverse class of biomolecules, including fatty acids (e.g., long-chain carbohydrates with carboxylate tail groups), diglycerides, triglycerides, and polyglycerides, phospholipids, prenols, sterols (e.g., cholesterol), and ladderanes, among many other types. While lipids are primarily found in membranes, free lipids, protein-complexed lipids, and nucleic acid-complexed lipids are typically present in various biological fluids and can, in some cases, be fractionated separately from membrane-bound lipids. For example, lipid-bound proteins (e.g., albumin) can be collected from samples by immunohistochemical precipitation and chemically induced to release bound lipids for subsequent collection and detection.

[0269] Lipids may be essential components in the development of diseases such as cancer. For example, lipids may play an important role in cancer biology by affecting or being involved in membrane and cell proliferation, lipotoxicity (where a balance of lipid content can help protect against lipotoxicity), enhancing cellular processes, membrane biophysics, oncogenic signaling and metastasis, protection from oxidative stress, signaling in the microenvironment, or immune regulation. Several lipid classes may be associated with cancer, such as glycerophospholipids in hepatocellular carcinoma, glycerophospholipids and acylcarnitines in prostate cancer, choline-containing lipids and phospholipids that increase during metastasis, or sphingolipids that regulate cancer cell survival and death.

[0270] Lipid data can be generated from a sample after processing the sample to isolate or enrich lipids in the sample. Generation of lipid data can include concentrating, filtering, or centrifuging the sample. Lipid analysis can include lipid fractionation. In many cases, lipids can be easily separated from other types of biomolecules for lipid-specific analysis. Because many lipids are highly hydrophobic, organic solvent extraction and gradient chromatography methods can cleanly separate lipids from other types of biomolecules present in the sample. Lipid data can be generated using mass spectrometry. Lipid analysis can then distinguish lipids by class (e.g., sphingolipids versus chlorolipids) or by individual type.

[0271] Lipidomic data can be generated by any of a variety of methods. The generation of lipidomic data can include the use of a detection reagent that binds to lipids and generates a detectable signal. After using the detection reagent that binds to lipids and generates a detectable signal, a reading can be obtained that indicates the presence, absence, or amount of lipid. The generation of lipidomic data can include concentrating, filtering, or centrifuging the sample.

[0272] Lipidomic data can be generated using mass spectrometry, chromatography, liquid chromatography, high-performance liquid chromatography, solid-phase chromatography, lateral flow assays, immunoassays, enzyme-linked immunosorbent assays, Western blots, dot blots, or immunostaining, or a combination thereof. One example of a method for generating lipidomic data includes using mass spectrometry. Mass spectrometry may include a separation method step, such as liquid chromatography (e.g., HPLC). Mass spectrometry may include an ionization method, such as electron ionization, atmospheric pressure chemical ionization, electrospray ionization, or secondary electrospray ionization. Mass spectrometry may include surface-based mass spectrometry or secondary ion mass spectrometry. Another example of a method for generating lipidomic data includes nuclear magnetic resonance (NMR). Other examples of methods for generating lipidomic data include Fourier transform ion cyclotron resonance, ion mobility spectrometry, electrochemical detection (e.g., coupled with HPLC), or Raman spectroscopy and radiolabeling (e.g., when combined with thin-layer chromatography). Some mass spectrometry methods described for generating lipidomics data can also be used to generate proteomics data, and vice versa. Lipidomic data can also be generated using immunoassays, such as enzyme-linked immunosorbent assays, Western blots, dot blots, or immunohistochemistry. Generating lipidomics data can also include the use of lipid panels.

[0273] In addition to any of the above methods, generating lipidomics data may include contacting the sample with particles such that the particles adsorb biomolecules, including lipids. The adsorbed lipids may be part of a biomolecular corona. The adsorbed lipids may be measured or identified in generating the lipidomics data.

[0274] The generation of lipidomic data may include the use of a known amount of an internal reference lipid. The reference lipid may be labeled. The label may include an isotopic label. The generation of lipidomic data may include the use of a known amount of an isotopically labeled internal reference lipid. The internal reference lipid may be spiked into a sample. The internal reference lipid can be used to identify the mass spectrum of each endogenous lipid. The internal reference lipid can be used as a standard for determining the amount of each endogenous lipid. The lipidomic measurements can be generated based on the amount of lipid added to one of one or more biological fluid samples. The lipidomic measurements can be generated based on the amount of labeled lipid added to one of one or more biological fluid samples.

[0275] Lipids may be associated with the biology of diseases such as cancer. Lipids may include phospholipids. Examples of phospholipids include phosphatidylethanolamine (PE), phosphatidylcholine (PC), phosphatidylinositol (PI), or phosphatidylglycerol (PG). Some phospholipids are components of cell membranes and may play cellular roles such as chemical energy storage, cell signaling, cell membranes, or cell interactions within tissues. Lipids may include ceramides (CER). Ceramides may act as tumor suppressors and may be a targeted therapeutic pathway. For example, the effectiveness of some chemotherapeutics and targeted therapies may depend on ceramide levels. Lipids may include diacylglycerides (DAG). Lipids may include triacylglycerides (TAG). Lipids may include fatty acids (FA).

[0276] Examples of lipids include PC(20:3_20:3)+AcO, Cer(d18:1 / 24:0)+H, GlcCer(d18:1 / 18:0+H, PI(18:0_18:3)-H, Aca(4:0)+H, GlcCer(d18:1 / 22:0+H, PC(18:2_20:5)+AcO, PC(14:0_18:2)+AcO, LPE(18:3)-H, Cer(d18:0 / 18:0)+H, DAG(18:1_22:6)+NH4, TAG(54:3_16:0)+NH4, Cer(d18:1 / The lipid data can include measurements of 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, or 20 of these lipids.

[0277] Example lipids may include any of the lipids in Figure 27. The lipid data may include measurements of 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, or 20 of these lipids, or any range of the aforementioned number of lipids from these figures.

[0278] Examples of lipids are shown in Figures 33A-33B. Lipids detected using the methods described herein may include CER(d18:1_10:0). Some examples of lipids are shown in Figure 36. Lipids detected using the methods described herein may include CER(d18.1_18.0), PC(18.2_20.5), CER(d18.1_24.1), CER(d18.1_16.0), TAG(56.5)_FA18.0), CER(d18.0_24.1), TAG(56.5_FA18.1), DAG(16.0_22.5), CER(d18.1_22.1), PE(P-18.0_18.3), or PE(17.0_22.6). Any number of the aforementioned lipids can be used. Any of the lipids can be used in the classifier.

[0279] Lipid measurements may be affected (e.g., decreased) in samples from subjects with liver cancer compared to lipid measurements from control samples or compared to baseline measurements. Lipid measurements may include phospholipid measurements. Lipid measurements may be useful in assessing liver cancer. Lipid measurements may include measurements of lipids or phospholipids, or combinations of lipids or phospholipids, from Figure 39F or Figure 39G. Lipid measurements may be useful in assessing ovarian cancer. Lipid measurements may include measurements of lipids or phospholipids, or combinations of lipids or phospholipids, from Figure 39F or Figure 40f.Lipid measurements included the following lipids: LPC.14.0 AcO, LPC.15.0 AcO, LPC.16.0 AcO, LPC.16.1 AcO, LPC.17.0 AcO, LPC.18.0 AcO, LPC.18.1 AcO, LPC.18.2 AcO, LPC.18.3 AcO, LPC.20.2 AcO, LPC.20.3 AcO, LPC.20.4 AcO, LPE.18.0 H, LPE.18.2 H, LPE.20.4 H, PA.18.0_18.2 H, PC.14.0_18 .2..AcO, PC.14.0_18.3..AcO, PC.14.0_20.2..AcO, PC.14.0_20.3..AcO, PC.14.0_20.4..AcO, PC.14.0_22.5..AcO, PC.14.0_22.6..AcO, PC.15. 0_18.2..AcO, PC.15.0_20.3..AcO, PC.15.0_20.4..AcO, PC.16.0_20.3..AcO, PC.16.1_20.3..AcO, PC.16.1_20.4..AcO, PC.18.0_18.2..AcO, PC. 18.0_20.3..AcO, PC.18.1_20.3..AcO, PC.18.1_20.4..AcO, PC.18.1_22.4..AcO, PC.18.1_22.5..AcO, PC.18.2_18.2..AcO, PC.18.2_18.3..AcO , PC.18.2_20.3..AcO, PC.18.2_20.4..AcO, PC.18.2_20.5..AcO, PC.20.2_20.3..AcO, PC.20.2_20.4..AcO, PC.20.3_20.3..AcO, PC.20.3_20.4.. The measurement may include one or more measurements of AcO, PC.20.4_20.4..AcO, PC.20.4_22.5..AcO, PE.O.16.0_20.3..H, PE.O.16.0_20.4..H, PE.O.16.0_22.5..H, or PI.18.1_20.4..H, where "LPC" indicates lysophosphatidylcholine, "LPE" indicates lysophosphatidylethanolamine, "PA" indicates phosphatidic acid, "PC" indicates phosphatidylcholine, and "PE" indicates phosphatidylethanolamine.A lipid or phospholipid combination may include 2, 3, 4, 5, 6, 7, 8, 9, 10, 15, 20, 25, 30, 35, 40, 45, or 50 lipids in Figure 39F, or a range defined by any two of the foregoing integers. A lipid or phospholipid combination may include at least 1, at least 2, at least 3, at least 4, at least 5, at least 6, at least 7, at least 8, at least 9, at least 10, at least 15, at least 20, at least 25, at least 30, at least 35, at least 40, or at least 45 lipids in Figure 39F. A lipid or phospholipid combination may include fewer than 3, fewer than 4, fewer than 5, fewer than 6, fewer than 7, fewer than 8, fewer than 9, fewer than 10, fewer than 15, fewer than 20, fewer than 25, fewer than 30, fewer than 35, fewer than 40, fewer than 45, or fewer than 50 lipids in Figure 39F. In some embodiments, the lipid combination does not include any one or more of the lipids in Figure 39F or Figure 39G. In some embodiments, the lipid combination does not include any one or more of the lipids in Figure 39F or Figure 40f.

[0280] Any of the following lipids may be useful in assessing ovarian cancer: LPC.14.0..AcO, LPC.15.0..AcO, LPC.16.0..AcO, LPC.16.1..AcO, LPC.17.0..AcO, LPC.18.0..AcO, LPC.18.1..AcO, LPC.18.2..AcO, LPC.18.3..AcO, LPC.20.2..AcO, LPC.20.3..AcO, or LPC.20.4..AcO.

[0281] Any of the following lipids: LPC.14.0..AcO, LPC.15.0..AcO, LPC.16.0..AcO, LPC.16.1..AcO, LPC.17.0..AcO, LPC.18.0..AcO, LPC.18.1..Ac O, LPC.18.2..AcO, LPC.18.3..AcO, LPC.20.2..AcO, LPC.20.3..AcO, LPC.20.4..AcO, LPE.18.0..H, LPE.18.2..H, LPE.20.4..H, PA.18.0_18.2..H, PC.14.0_18.2..AcO, PC.14.0_18.3..AcO, PC.14.0_20.2..AcO, PC.14.0_20.3..AcO, PC.14.0_20.4..AcO, P C.14.0_22.5..AcO, PC.14.0_22.6..AcO, PC.15.0_18.2..AcO, PC.15.0_20.3..AcO, PC.15.0_20.4..AcO, PC.16.0_20.3..AcO, P C.16.1_20.3..AcO, PC.16.1_20.4..AcO, PC.18.0_18.2..AcO, PC.18.0_20.3..AcO, PC.18.1_20.3..AcO, PC.18.1_20.4..AcO, PC.18.1_22.4..AcO, PC.18.1_22.5..AcO, PC.18.2_18.2..AcO, PC.18.2_18.3..AcO, PC.18.2_20.3..AcO, PC.18.2_20.4..AcO, PC.18.2_20.5..AcO, PC.20.2_20.3..AcO, PC.20.2_20.4..AcO, PC.20.3_20.3..AcO, PC.20.3_20.4..AcO, PC.20.4_20.4..AcO, PC.20.4_22.5..AcO, PE.O.16.0_20.3..H, PE.O.16.0_204..H, PE.O.16.0_22.5..H, or PI.18.1_20.4..H may be useful in the evaluation of liver cancer.

[0282] Metabolomics Data Data such as the multi-omics data described herein may include metabolite data or metabolomics data. Metabolomics data may include information about small molecule (e.g., less than 1.5 kDa) metabolites (such as metabolic intermediates, hormones or other signaling molecules, or secondary metabolites). Metabolomics data may include data about metabolites. Metabolites may include metabolic substrates, intermediates, or products. Metabolites may include small molecules. Metabolites may be any molecule less than 1.5 kDa in size. Examples of metabolites may include sugars, lipids, amino acids, fatty acids, phenolic compounds, or alkaloids. Metabolomics data can be distinguished by subtypes, with each subtype containing different types of metabolite data. Metabolomics data may include some lipid data. Metabolomics data may include lipidomics data. Aspects described with respect to metabolomics data may also be relevant to metabolite data, and vice versa. Metabolomics data may include measurements of metabolites. Metabolite measurements may include measurements of lipids, such as phospholipids.

[0283] Metabolomics data may include information regarding the presence, absence, or amount of various metabolites. For example, metabolomics data may include the amount of a metabolite. The amount of a metabolite may be expressed as the concentration or amount of the metabolite, for example, the concentration of the metabolite in a biological fluid. The amount of a metabolite may be compared to another metabolite or another biomolecule. Metabolomics data may include information regarding the presence of a metabolite. Metabolomics data may include information regarding the absence of a metabolite.

[0284] Metabolomics data generally includes data on a large number of metabolites. For example, metabolomics data may include information on the presence, absence, or amount of 1,000 or more metabolites. In some cases, metabolomics data may include information on the presence, absence, or amount of 5,000, 10,000, 20,000, 50,000, 100,000, 500,000, 1 million, 1.5 million, 2 million, or more metabolites, or a range defined by any two of the aforementioned numbers of metabolites.

[0285] Metabolomic data can be generated by any of a variety of methods. Generating metabolomic data can include using a detection reagent that binds to a metabolite and generates a detectable signal. After using a detection reagent that binds to a metabolite and generates a detectable signal, a readout can be obtained that indicates the presence, absence, or amount of the metabolite. Generating metabolomic data can include concentrating, filtering, or centrifuging the sample.

[0286] Metabolomic data can be generated using mass spectrometry, chromatography, liquid chromatography, high-performance liquid chromatography, solid-phase chromatography, lateral flow assays, immunoassays, enzyme-linked immunosorbent assays, Western blots, dot blots, or immunostaining, or a combination thereof. One example of a method for generating metabolomic data includes using mass spectrometry. Mass spectrometry may include a separation method step, such as liquid chromatography (e.g., HPLC). Mass spectrometry may include an ionization method, such as electron ionization, atmospheric pressure chemical ionization, electrospray ionization, or secondary electrospray ionization. Mass spectrometry may include surface-based mass spectrometry or secondary ion mass spectrometry. Another example of a method for generating metabolomic data includes nuclear magnetic resonance (NMR). Other examples of methods for generating metabolomic data include Fourier transform ion cyclotron resonance, ion mobility spectrometry, electrochemical detection (e.g., coupled with HPLC), or Raman spectroscopy and radiolabeling (e.g., when combined with thin-layer chromatography). Some mass spectrometry methods described for generating metabolomic data can also be used to generate proteomic data, and vice versa. Metabolomic data can also be generated using immunoassays, such as enzyme-linked immunosorbent assays, Western blots, dot blots, or immunohistochemistry. Generating metabolomic data can also include the use of lipid panels.

[0287] In addition to any of the above methods, generating metabolomics data may include contacting the sample with particles such that the particles adsorb biomolecules, including metabolites. The adsorbed metabolites may be part of a biomolecular corona. The adsorbed metabolites may be measured or identified in generating the metabolomics data.

[0288] The generation of metabolomics data may include the use of a known amount of an internal reference metabolite. The reference metabolite may be labeled. The label may include an isotopic label. The generation of metabolomics data may include the use of a known amount of an isotopically labeled internal reference metabolite. The internal reference metabolite may be spiked into a sample. The internal reference metabolite can be used to identify the mass spectrum of each endogenous metabolite. The internal reference metabolite can be used as a standard for determining the amount of each endogenous metabolite. The metabolomics measurements can be generated based on the amount of a metabolite added to one of one or more biological fluid samples. The metabolomics measurements can be generated based on the amount of a labeled metabolite added to one of one or more biological fluid samples.

[0289] Examples of metabolites are shown in Figures 34A-34B. Metabolites detected by the methods described herein can include 5-aminoimidazole-4-carboxamide ribonucleotide (AICAR). Metabolites can include nucleotides, such as monophosphate nucleotides. Some examples of metabolites are shown in Figure 36. Metabolites detected by the methods described herein can include cytidine monophosphate (CMP). Metabolites can include AICAR or CMP. Detected metabolites can include AICAR and CMP. Any number of the aforementioned metabolites can be used. Any metabolite can be used.

[0290] Using Particles The sample can be contacted with particles, for example, before generating data. The data described herein can be generated using particles. For example, a method can include contacting a sample with particles such that the particles adsorb biomolecules. The particles can attract a set of biomolecules that are different from those that would normally be accurately measured by performing omics measurements directly on the sample. For example, a dominant biomolecule may constitute a significant percentage of a particular type of biomolecules (e.g., proteins, transcripts, genetic material, or metabolites) in a sample. For example, the majority of circulating proteins collected by blood sampling may consist of a single protein. By attaching biomolecules to particles before analyzing the biomolecules, a subset of biomolecules that does not contain the dominant biomolecule can be obtained. Removing the dominant biomolecule in this manner can improve the accuracy of biomolecule measurements and the sensitivity of analyses using those measurements.

[0291] Examples of biomolecules that can be adsorbed to particles include proteins, transcripts, genetic material, or metabolites. The adsorbed biomolecules may form a biomolecular corona around the particle. The adsorbed biomolecules can be measured or identified in generating data, such as omics data (e.g., proteomics data). In some embodiments, proteomics measurements are generated from proteins adsorbed to the nanoparticles. The nanoparticles can enrich for proteins or other types of biomolecules.

[0292] Particles can be made from a variety of materials. Such materials may include metals, magnetic particles, polymers, or lipids. Particles can be made from a combination of materials. Particles may include layers of different materials. Different materials may have different properties. Particles may include a core comprising one material and coated with another material. The core and coating may have different properties.

[0293] The particles can comprise a metal, for example, the particles can comprise gold, silver, copper, nickel, cobalt, palladium, platinum, iridium, osmium, rhodium, ruthenium, rhenium, vanadium, chromium, manganese, niobium, molybdenum, tungsten, tantalum, iron, or cadmium, or a combination thereof.

[0294] The particles may be magnetic (e.g., ferromagnetic or ferrimagnetic). The particles may be magnetic, including iron oxide. The particles may include superparamagnetic iron oxide nanoparticles (SPIONs).

[0295] The particle may contain a polymer. Examples of polymers include polyethylene, polycarbonate, polyanhydride, polyhydroxy acid, polypropyl fumerate, polycaprolactone, polyamide, polyacetal, polyether, polyester, poly(orthoester), polycyanoacrylate, polyvinyl alcohol, polyurethane, polyphosphazene, polyacrylate, polymethacrylate, polycyanoacrylate, polyurea, polystyrene, or polyamine, polyalkylene glycol (e.g., polyethylene glycol (PEG)), polyester (e.g., poly(lactide-co-glycolide) (PLGA), polylactic acid, or polycaprolactone), or a copolymer of two or more polymers, for example, a copolymer of polyalkylene glycol (e.g., PEG) and polyester (e.g., PLGA). The particle can be made from a combination of polymers.

[0296] The particles may comprise lipids. Examples of lipids include dioleoylphosphatidylglycerol (DOPG), diacylphosphatidylcholine, diacylphosphatidylethanolamine, ceramide, sphingomyelin, cephalin, cholesterol, cerebrosides, and diacylglycerol, dioleoylphosphatidylcholine (DOPC), dimyristoylphosphatidylcholine (DMPC), and dioleoylphosphatidylserine (DOPS), phosphatidylglycerol, cardiolipin, diacylphosphatidylserine, diacylphosphatidic acid, N-dodecanoylphosphatidylethanolamine, N-succinylphosphatidylethanolamine, N-glutarylphosphatidylethanolamine, lysylphosphatidylglycerol, palmitoyloleoylphosphatidylglycerol (POPG), lecithin, lysolecithin, phosphatidylethanolamine, lysophosphatides, and the like. Dioleoyl phosphatidylethanolamine, dioleoyl phosphatidylethanolamine (DOPE), dipalmitoyl phosphatidylethanolamine (DPPE), dimyristoylphosphoethanolamine (DMPE), distearoyl phosphatidylethanolamine (DSPE), palmitoyloleoyl phosphatidylethanolamine (POPE), palmitoyloleoyl phosphatidylcholine (POPC), egg yolk phosphatidylcholine (EPC), distearoyl phosphatidylcholine (DSPC), dioleoyl phosphatidylcholine (DOPC), dipalmitoyl phosphatidylcholine (DPPC), dioleoyl phosphatidylglycerol (DOPG), dipalmitoyl phosphatidylglycerol (DPPG), palmitoyloleoyl phosphatidylglycerol (POPG), 16-O-monomethyl PE, 16-O-dimethyl PE, 18-1-trans Examples of lipids include PE, palmitoyloleoyl-phosphatidylethanolamine (POPE), 1-stearoyl-2-oleoyl-phosphatidylethanolamine (SOPE), phosphatidylserine, phosphatidylinositol, sphingomyelin, cephalin, cardiolipin, phosphatidic acid, cerebroside, dicetyl phosphate, or cholesterol. Particles can be made from a combination of lipids.

[0297] Further examples of materials include silica, carbon, carboxylate, polyacrylic acid, carbohydrate, dextran, polystyrene, dimethylamine, amine, or silane. Some example particles include carboxylate SPIONs, phenol-formaldehyde coated SPIONs, silica coated SPIONs, polystyrene coated SPIONs, carboxylated poly(styrene-co-methacrylic acid), P(St-co-MAA) coated SPIONs, N-(3-trimethoxysilylpropyl)diethylenetriamine coated SPIONs, poly(N-(3-(dimethylamino)propyl)methacrylamide) (PDMAPMA) coated SPIONs, 1,2,4,5-benzenetetracarboxylic acid coated SPIONs, poly(vinylbenzyltrimethylammonium chloride) (PVBTMAC) coated SPIONs, carboxylates coated with peracetic acid, poly(oligo(ethylene glycol)methyl ether methacrylate) (POEGMA) coated SPIONs, polystyrene carboxyl functionalized particles, carboxylic acid particles, particles with an amino surface, silica amino functionalized particles, particles with a Jeffamine surface, or silica silanol coated particles.

[0298] Particles of various sizes can be used. The particles may include nanoparticles. The nanoparticles may have a diameter of about 10 nm to about 1000 nm. For example, the nanoparticles may have a diameter of at least 10 nm, at least 100 nm, at least 200 nm, at least 300 nm, at least 400 nm, at least 500 nm, at least 600 nm, at least 700 nm, at least 800 nm, at least 900 nm, 10 nm to 50 nm, 50 nm to 100 nm, 100 nm to 150 nm, 150 nm to 200 nm, 200 nm to 250 nm, 250 nm to 300 nm, 300 nm to 350 nm, 350 nm to 400 nm, 400 nm to 450 nm, 450 nm to 500 nm, 500 nm to 550 nm, 5 ... The diameter of the nanoparticles may be less than 1000 nm. Some examples include diameters of about 50 nm, about 130 nm, about 150 nm, 400-600 nm, or 100-390 nm.

[0299] The particles may include microparticles. The microparticles may be particles having a diameter of about 1 μm to about 1000 μm. For example, the microparticles may have a diameter of at least 1 μm, at least 10 μm, at least 100 μm, at least 200 μm, at least 300 μm, at least 400 μm, at least 500 μm, at least 600 μm, at least 700 μm, at least 800 μm, at least 900 μm, 10 μm to 50 μm, 50 μm to 100 μm, 100 μm to 150 μm, 150 μm to 200 μm, 200 μm to 250 μm, 250 μm to 300 μm, 300 μm to 350 μm, 350 μm to 400 μm, 400 μm to 450 μm, 450 μm to 500 μm, 500 μm to 550 μm, 550μm~600μm, 600μm~650μm, 650μm~700μm, 700μm~750μm, 750μm~800μm, 800μm m~850μm, 850μm~900μm, 100μm~300μm, 150μm~350μm, 200μm~400μm, 250μm~450 The diameter of the microparticles may be less than 1000 μm. Some examples include diameters of 2.0 to 2.9 μm.

[0300] The particles can include physiochemically distinct sets of particles (e.g., two or more sets of physiochemical particles, where one set of particles is physiochemically distinct from another set of particles). Examples of physiochemical properties include charge (e.g., positive, negative, or neutral) or hydrophobicity (e.g., hydrophobic or hydrophilic). The particles can include 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, or more particle sets, or a range of particle sets including any of the foregoing numbers of particle sets.

[0301] Particles and Types Disease detection methods can include the use of particles. The methods described herein can include contacting a biological sample with physiochemically distinct particles to form a biomolecular corona. The biological sample can be from a subject identified as having a pulmonary nodule. The particles can adsorb biomolecules from the biological sample, thereby forming a biomolecular corona on the particle's surface. Upon contact with the biological sample, the particles can adsorb multiple peptides, proteins, nucleic acids, lipids, sugars, small molecules (such as metabolites (natural and foreign), terpenes, polyketides, and cyclic peptides), or any combination thereof. Thus, methods include collecting a subset of biomolecules from a biological sample (e.g., a complex biological sample such as human plasma) onto particles, analyzing the biomolecules collected on the particles, analyzing the biomolecules remaining in the biological sample, or analyzing the biomolecules collected on the particles and the biomolecules remaining in the biological sample. The biomolecules, biomolecular corona, or portions thereof can be eluted from the particles into solution prior to analysis. In some embodiments, assaying for proteins includes contacting a biological fluid sample with particles such that the particles adsorb proteins to the particles.

[0302] The relationship between particle properties and biomolecular corona composition can be exploited to manipulate biomolecule collection from samples. In some cases, a set of particle properties may favor the binding of specific biomolecule types, families, or superfamilies. For example, humans express over 100 proteins from the Ras superfamily that share a conserved GTP-binding motif within their 20 kilodalton (kDa) N-terminal domain. Particles or particle assemblies (e.g., a mixture containing five types of particles) can be functionalized to promote Ras protein adsorption and thus tailored to preferentially adsorb Ras proteins from complex biological samples, potentially enabling their enrichment for further analysis.

[0303] Particles or mixtures of different particles can be tailored to perform broad profiling of samples. In many biological samples, a small number of biomolecules constitute the majority of the biological material. For example, over 99% of the protein mass in human plasma is accounted for by only 20 of the approximately 3,500 human plasma proteins. Analysis of such samples can be very challenging because a small number of highly abundant biomolecules can saturate detection or enrichment schemes. Particles or collections of multiple particle types can be tailored to perform broad profiling of complex biological material, such that low-abundance biomolecules from a complex biological sample are preferentially enriched over or with higher-abundance biomolecules. Particles or collections of multiple particle types have similar binding affinities for multiple biomolecules, thereby facilitating the adsorption of multiple biomolecules from a sample. Particles may have low affinity for a highly abundant protein or set of highly abundant proteins in a sample, thereby preferentially adsorbing and enriching low-abundance biomolecules. Particle collections can include particle types with affinities for different types or classes of biomolecules, such that the particle collection adsorbs a wide range of biomolecules from a sample. Thus, the present disclosure provides a wide range of particle types with different physicochemical properties.

[0304] Particle types consistent with the methods disclosed herein can be made from a variety of materials. For example, particle materials consistent with the present disclosure include metals, polymers, magnetic materials, and lipids. Magnetic particles may be iron oxide particles. Examples of metallic materials include gold, silver, copper, nickel, cobalt, palladium, platinum, iridium, osmium, rhodium, ruthenium, rhenium, vanadium, chromium, manganese, niobium, molybdenum, tungsten, tantalum, iron, and cadmium, or any other material described in U.S. Pat. No. 7,749,299, the contents of which are incorporated herein by reference in their entirety. Particles may be magnetic (e.g., ferromagnetic or ferrimagnetic). For example, particles may include superparamagnetic iron oxide nanoparticles (SPIONs).

[0305] The particles may comprise a plurality of physiochemically distinct particles (e.g., two or more sets of physiochemically distinct particles, where one set of particles is physiochemically distinct from another set of particles). In some embodiments, the particles comprise nanoparticles. In some embodiments, the particles comprise a population of physiochemically distinct nanoparticles. The physiochemically distinct particles may comprise lipid particles, metal particles, silica particles, or polymer particles. The physiochemically distinct particles may comprise carboxylate particles, polyacrylic acid particles, dextran particles, polystyrene particles, dimethylamine particles, amino particles, silica particles, or N-(3-trimethoxysilylpropyl)diethylenetriamine particles.

[0306] The particles may contain a polymer. Examples of polymers include polyethylene, polycarbonate, polyanhydride, polyhydroxy acid, polypropyl fumerate, polycaprolactone, polyamide, polyacetal, polyether, polyester, poly(orthoester), polycyanoacrylate, polyvinyl alcohol, polyurethane, polyphosphazene, polyacrylate, polymethacrylate, polycyanoacrylate, polyurea, polystyrene, or polyamine, polyalkylene glycol (e.g., polyethylene glycol (PEG)), polyester (e.g., poly(lactide-co-glycolide) (PLGA), polylactic acid, or polycaprolactone), or copolymers of two or more polymers, such as copolymers of polyalkylene glycol (e.g., PEG) and polyester (e.g., PLGA), or any combination thereof. The polymer may be lipid-terminated polyalkylene glycol and polyester, or any other material disclosed in US9549901, the contents of which are incorporated herein by reference in their entirety.

[0307] The particles may comprise lipids. Examples of lipids that can be used to form the particles of the present disclosure include cationic, anionic, and neutrally charged lipids. For example, the particles may be composed of dioleoylphosphatidylglycerol (DOPG), diacylphosphatidylcholine, diacylphosphatidylethanolamine, ceramide, sphingomyelin, cephalin, cholesterol, cerebrosides and diacylglycerol, dioleoylphosphatidylcholine (DOPC), dimyristoylphosphatidylcholine (DMPC), and dioleoylphosphatidylserine (DOPS), phosphatidylglycerol, cardiolipin, diacylphosphatidylserine, diacylphosphatidic acid, N-dodecanoylphosphatidylethanolamine, N-succinylphosphatidylethanolamine, N-glutarylphosphatidylethanolamine, lysylphosphatidylglycerol, palmitoyloleoylphosphatidylglycerol (POPG), lecithin, lysolecithin, phosphatidylethanolamine, lysophosphatidyl Dioleoyl phosphatidylethanolamine, dioleoyl phosphatidylethanolamine (DOPE), dipalmitoyl phosphatidylethanolamine (DPPE), dimyristoylphosphoethanolamine (DMPE), distearoyl phosphatidylethanolamine (DSPE), palmitoyloleoyl phosphatidylethanolamine (POPE), palmitoyloleoyl phosphatidylcholine (POPC), egg yolk phosphatidylcholine (EPC), distearoyl phosphatidylcholine (DSPC), dioleoyl phosphatidylcholine (DOPC), dipalmitoyl phosphatidylcholine (DPPC), dioleoyl phosphatidylglycerol (DOPG), dipalmitoyl phosphatidylglycerol (DPPG), palmitoyloleoyl phosphatidylglycerol (POPG), 16-O-monomethyl PE, 16-O-dimethyl PE, 18-1-transThe particles can be made from any one or any combination of PE, palmitoyloleoyl-phosphatidylethanolamine (POPE), 1-stearoyl-2-oleoyl-phosphatidylethanolamine (SOPE), phosphatidylserine, phosphatidylinositol, sphingomyelin, cephalin, cardiolipin, phosphatidic acid, cerebrosides, dicetyl phosphate, and cholesterol, or any other materials disclosed in U.S. Pat. No. 9,445,994, the contents of which are incorporated herein by reference in their entirety. Examples of particles of the present disclosure are provided in Table 1. [Table 1]

[0308] Examples of particle types of the present disclosure include carboxylate (citrate) superparamagnetic iron oxide nanoparticles (SPIONs), phenol formaldehyde coated SPIONs, silica coated SPIONs, polystyrene coated SPIONs, carboxylated poly(styrene-co-methacrylic acid) coated SPIONs, N-(3-trimethoxysilylpropyl)diethylenetriamine coated SPIONs, poly(N-(3-(dimethylamino)propyl)methacrylamide) (PDMAPMA) coated SPIONs, 1,2,4,5-benzenetetracarboxylic acid coated SPIONs, poly(vinylbenzyltrimethylammonium chloride) (PVBTMAC) coated SPIONs, carboxylate, PAA coated SPIONs, and poly(styrene-co-methacrylic acid) coated SPIONs. The microparticles can be coated SPIONs, poly(oligo(ethylene glycol) methyl ether methacrylate) (POEGMA)-coated SPIONs, carboxylate microparticles, polystyrene carboxyl-functionalized particles, carboxylic acid-coated particles, silica particles, carboxylic acid particles with a diameter of approximately 150 nm, amino-surfaced microparticles with a diameter of approximately 0.4-0.6 μm, silica amino-functionalized microparticles with a diameter of approximately 0.1-0.39 μm, Jeffamine-surfaced particles with a diameter of approximately 0.1-0.39 μm, polystyrene microparticles with a diameter of approximately 2.0-2.9 μm, silica particles, carboxylated particles with an original coating with a diameter of approximately 50 nm, particles coated with a dextran-based coating with a diameter of approximately 0.13 μm, or silica silanol-coated particles with low acidity.

[0309] Particles consistent with the present disclosure can be produced and used in methods to form protein coronas in a wide range of sizes after incubation in biological fluids. The particles of the present disclosure can be nanoparticles. The nanoparticles of the present disclosure can have diameters ranging from about 10 nm to about 1000 nm. For example, the nanoparticles of the present disclosure can have diameters of at least 10 nm, at least 100 nm, at least 200 nm, at least 300 nm, at least 400 nm, at least 500 nm, at least 600 nm, at least 700 nm, at least 800 nm, at least 900 nm, 10 nm to 50 nm, 50 nm to 100 nm, 100 nm to 150 nm, 150 nm to 200 nm, 200 nm to 250 nm, 250 nm to 300 nm, 300 nm to 350 nm, 350 nm to 400 nm, 400 nm to 450 nm, 450 nm to 500 nm, 500 nm to 550 nm, 550 nm to 600 nm, 600 nm to 700 nm, 700 nm to 800 nm, 800 nm to 900 nm, 800 nm to 900 nm, 900 nm to 1000 nm, 100 nm to 150 nm, 150 nm to 200 nm, 200 nm to 250 nm, 250 nm to 300 nm, 300 nm to 350 nm, 350 nm to 400 nm, 400 nm to 450 nm, 450 nm to 500 nm, 500 nm to 550 nm, 550 nm to 600 nm, 600 nm to 700 nm, 700 nm The diameter of the nanoparticles may be less than 1000 nm.

[0310] The particles of the present disclosure may be microparticles. The microparticles may be particles having a diameter of about 1 μm to about 1000 μm. For example, the microparticles of the present disclosure may have a diameter of at least 1 μm, at least 10 μm, at least 100 μm, at least 200 μm, at least 300 μm, at least 400 μm, at least 500 μm, at least 600 μm, at least 700 μm, at least 800 μm, at least 900 μm, 10 μm to 50 μm, 50 μm to 100 μm, 100 μm to 150 μm, 150 μm to 200 μm, 200 μm to 250 μm, 250 μm to 300 μm, 300 μm to 350 μm, 350 μm to 400 μm, 400 μm to 450 μm, 450 μm to 500 μm, or 500 μm to 550 μm. m, 550μm~600μm, 600μm~650μm, 650μm~700μm, 700μm~750μm, 750μm~800μm, 800 μm~850μm, 850μm~900μm, 100μm~300μm, 150μm~350μm, 200μm~400μm, 250μm~450 The diameter of the microparticles may be less than 1000 μm.

[0311] The surface area to mass ratio can be a determining factor in particle properties. For example, the number and type of biomolecules that a particle will adsorb from solution will vary depending on the surface area to mass ratio of the particle. The particles disclosed herein have a surface area of ​​3-30 cm 2 / mg, 5-50cm 2 / mg, 10-60cm 2 / mg, 15-70cm 2 / mg, 20-80cm 2 / mg, 30-100cm 2 / mg, 35-120cm 2 / mg, 40-130cm 2 / mg, 45-150cm 2 / mg, 50-160cm 2 / mg, 60-180cm 2 / mg, 70-200cm2 / mg, 80-220cm 2 / mg, 90-240cm 2 / mg, 100-270cm 2 / mg, 120-300cm 2 / mg, 200-500cm 2 / mg, 10-300cm 2 / mg, 1-3000cm 2 / mg, 20-150cm 2 / mg, 25-120cm 2 / mg, or 40-85cm 2 / mg. Small particles (e.g., 50 nm or less in diameter) can have significantly larger surface area to mass ratios, which is due in part to a higher order dependence of diameter on mass rather than surface area. In some cases (e.g., for small particles), particles can have a surface area to mass ratio of 200-1000 cm 2 / mg, 500-2000cm 2 / mg, 1000-4000cm 2 / mg, 2000-8000cm 2 / mg, or 4000-10000cm 2 In some cases (e.g., for larger particles), the particles may have a surface area to mass ratio of 1-3 cm 2 / mg, 0.5-2cm 2 / mg, 0.25-1.5cm 2 / mg, or 0.1-1cm 2 / mg surface area to mass ratio.

[0312] In some cases, the plurality of particles (e.g., of a particle panel) used in the methods described herein may have a range of surface area to mass ratios. In some cases, the range of surface area to mass ratios of the plurality of particles is greater than or equal to 100 cm. 2 / mg, 80cm 2 / mg, 60cm 2 / mg, 40cm 2 / mg, 20cm 2 / mg, 10cm 2 / mg, 5cm 2 / mg, or 2cm 2 / mg. In some cases, the surface area to mass ratio of the plurality of particles varies by no more than 40%, 30%, 20%, 10%, 5%, 3%, 2%, or 1% among the plurality of particles. In some cases, the plurality of particles can include at least 2, 3, 4, 5, 6, 7, 8, 9, 10, 12, 15, 20, or more different types of particles.

[0313] In some cases, the plurality of particles (e.g., of a particle panel) may have a wider range of surface area to mass ratios. In some cases, the surface area to mass ratio range of the plurality of particles is greater than 100 cm. 2 / mg more than 150cm 2 / mg, 200cm 2 / mg, 250cm 2 / mg, 300cm 2 / mg, 400cm 2 / mg, 500cm 2 / mg, 800cm 2 / mg, 1000cm 2 / mg, 1200cm 2 / mg, 1500cm 2 / mg, 2000cm 2 / mg, 3000cm 2 / mg, 5000cm 2 / mg, 7500cm 2 / mg, 10000cm 2 / mg, or more. In some cases, the surface area to mass ratio of the plurality of particles (e.g., within a panel) can vary by more than 100%, 200%, 300%, 400%, 500%, 1000%, 10,000% or more. In some cases, the plurality of particles having a wide range of surface area to mass ratios includes at least 2, 3, 4, 5, 6, 7, 8, 9, 10, 12, 15, 20, or more different types of particles.

[0314] The surface functional group may include a polymerizable functional group, a positively or negatively charged functional group, a zwitterionic functional group, an acidic or basic functional group, a polar functional group, or any combination thereof. The surface functional group may include a carboxyl group, a hydroxyl group, a thiol group, a cyano group, a nitro group, an ammonium group, an alkyl group, an imidazolium group, a sulfonium group, a pyridinium group, a pyrrolidinium group, a phosphonium group, an aminopropyl group, an amine group, a boronic acid group, an N-succinimidyl ester group, a PEG group, a streptavidin group, a methyl ether group, a triethoxypropylaminosilane group, a PCP group, a citric acid group, a lipoic acid group, a BPEI group, or any combination thereof. The particles of the plurality of particles may be selected from the group consisting of micelles, liposomes, iron oxide particles, silver particles, gold particles, palladium particles, quantum dots, platinum particles, titanium particles, silica particles, metal or inorganic oxide particles, synthetic polymer particles, copolymer particles, terpolymer particles, polymer particles with a metal core, polymer particles with a metal oxide core, polystyrene sulfonate particles, polyethylene oxide particles, polyoxyethylene glycol particles, polyethyleneimine particles, polylactic acid particles, polycaprolactone particles, polyglycolic acid particles, poly(lactide-co-glycolide) ... mer particles, cellulose ether polymer particles, polyvinylpyrrolidone particles, polyvinyl acetate particles, polyvinylpyrrolidone-vinyl acetate copolymer particles, polyvinyl alcohol particles, acrylate particles, polyacrylic acid particles, crotonic acid copolymer particles, polyethylene phosphonate particles, polyalkylene particles, carboxyvinyl polymer particles, sodium alginate particles, carrageenan particles, xanthan gum particles, acacia gum particles, gum arabic particles, guar gum particles, pullulan particles, agar particles, chitin particles, chitosan particles, pectin particles, karayatamtum) particles, locust bean gum particles, maltodextrin particles, amylose particles, corn starch particles, potato starch particles, rice starch particles, tapioca starch particles, pea starch particles, sweet potato starch particles, barley starch particles, wheat starch particles, hydroxypropylated high amylose starch particles, dextrin particles, levan particles, elsinan particles, gluten particles, collagen particles, isolated whey protein particles, casein particles, milk protein particles, soy protein particles, keratin particles, polyethylene particles, polycarbonate particles, polyanhydride particles, polyhydroxy acid particles, polypropylene fumarate particles, polycaprolactone particles, polyamine particles, polyacetal particles, polyether particles, poly polyester particles, poly(orthoester) particles, polycyanoacrylate particles, polyurethane particles, polyphosphazene particles, polyacrylate particles, polymethacrylate particles, polycyanoacrylate particles, polyurea particles, polyamine particles, polystyrene particles, poly(lysine) particles, chitosan particles, dextran particles, poly(acrylamide) particles, derivatized poly(acrylamide) particles, gelatin particles, starch particles, chitosan particles, dextran particles, gelatin particles, starch particles, poly-β-aminoester particles, poly(amidoamine) particles, polylactic-co-glycolic acid particles, polyanhydride particles, bioreducible polymer particles, and 2-(3-aminopropylamino)ethanol particles, and any combination thereof.

[0315] Multiple particles (e.g., physicochemically distinct particles) can be prepared using carboxylate (citrate) superparamagnetic iron oxide nanoparticles (SPIONs), phenol formaldehyde-coated SPIONs, silica-coated SPIONs, polystyrene-coated SPIONs, carboxylated poly(styrene-co-methacrylic acid)-coated SPIONs, N-(3-trimethoxysilylpropyl)diethylenetriamine-coated SPIONs, poly(N-(3-(dimethylamino)propyl)methacrylamide) (PDMAPMA)-coated SPIONs, 1,2,4,5-benzenetetracarboxylic acid-coated SPIONs, poly(vinylbenzyltrimethylamine)-coated SPIONs, and poly(vinylbenzyltrimethylamine). The particles may comprise one or more particle types selected from the group consisting of poly(trimethylammonium chloride) (PVB TMAC)-coated SPIONs, carboxylates, PAA-coated SPIONs, poly(oligo(ethylene glycol) methyl ether methacrylate) (POEGMA)-coated SPIONs, carboxylate microparticles, polystyrene carboxyl-functionalized particles, carboxylic acid-coated particles, silica particles, carboxylic acid particles, amino-surfaced particles, silica amino-functionalized particles, Jeffamine-surfaced particles, polystyrene particles, particles coated with a dextran-based coating having a diameter of about 0.13 μm, or silica silanol-coated particles.

[0316] Multiple particles (e.g., physicochemically distinct particles) can be prepared using carboxylate (citrate) superparamagnetic iron oxide nanoparticles (SPIONs), phenol formaldehyde-coated SPIONs, silica-coated SPIONs, polystyrene-coated SPIONs, carboxylated poly(styrene-co-methacrylic acid)-coated SPIONs, N-(3-trimethoxysilylpropyl)diethylenetriamine-coated SPIONs, poly(N-(3-(dimethylamino)propyl)methacrylamide) (PDMAPMA)-coated SPIONs, 1,2,4,5-benzenetetracarboxylic acid-coated SPIONs, poly(vinylbenzyltrimethylamine)-coated SPIONs, and poly(vinylbenzyltrimethylamine). The particles may comprise one or more particle types selected from the group consisting of poly(trimethylammonium chloride) (PVB TMAC)-coated SPIONs, carboxylates, PAA-coated SPIONs, poly(oligo(ethylene glycol) methyl ether methacrylate) (POEGMA)-coated SPIONs, carboxylate microparticles, polystyrene carboxyl-functionalized particles, carboxylic acid-coated particles, silica particles, carboxylic acid particles, amino-surfaced particles, silica amino-functionalized particles, Jeffamine-surfaced particles, polystyrene particles, particles coated with a dextran-based coating having a diameter of about 0.13 μm, or silica silanol-coated particles.

[0317] The plurality of particles (e.g., physicochemically distinct particles) may comprise one or more particle types selected from the group consisting of silica particles, poly(acrylamide) particles, polyethylene glycol particles, or combinations thereof. One or more of the particles may comprise a paramagnetic or superparamagnetic core material. The particles may comprise silica particles. The particles may comprise poly(acrylamide) particles. The particles may comprise polyethylene glycol particles.

[0318] The plurality of particles may comprise multiple particle types. In some cases, the plurality of particles comprises at least two types of particles. In some cases, the plurality of particles comprises at least three types of particles. In some cases, the plurality of particles comprises at least five types of particles. In some cases, the plurality of particles comprises at least six types of particles. In some cases, the plurality of particles comprises at least eight types of particles. In some cases, the plurality of particles comprises at least ten types of particles. In some cases, the plurality of particles comprises at least twelve types of particles. In some cases, the plurality of particles comprises at least fifteen types of particles. In some cases, the plurality of particles comprises at least eighteen types of particles. In some cases, the plurality of particles comprises at least twenty types of particles.

[0319] A particle may include layers with different properties. A particle may include a core with a first set of properties and a shell with a second set of properties. A particle may include multiple shells with different properties (e.g., a core including a first material, an inner shell including a second material, and an outer shell including a third material). A layer of a particle may include multiple materials. For example, a layer of a particle may include multiple polymers. The polymers may be homogeneously dispersed within the layer, phase separated, or applied heterogeneously.

[0320] In some cases, the one or more physicochemical properties are selected from the group consisting of composition, size, surface charge, hydrophobicity, hydrophilicity, surface functional groups, surface topography, surface curvature, shape, and any combination thereof. In some embodiments, the surface functional groups comprise chemical functionalization. In some embodiments, the small molecule functionalization comprises amine functionalization, carboxylate functionalization, monosaccharide functionalization, oligosaccharide functionalization, sugar phosphate functionalization, sugar sulfate functionalization, alcohol functionalization, ether functionalization, ester functionalization, amide functionalization, carbonate functionalization, carbamate functionalization, urea functionalization, benzyl functionalization, phenyl functionalization, phenol functionalization, aniline functionalization, imidazole functionalization, indole functionalization, fluoride functionalization, chloride functionalization, bromide functionalization, sulfide functionalization, nitro functionalization, thiol functionalization, nitrogen base functionalization, aminopropyl functionalization, boronic acid functionalization, N-succinimidyl ester functionalization, PEG functionalization, methyl ether functionalization, triethoxylpropylaminosilane functionalization, silicon alkoxide functionalization, phenol-formaldehyde functionalization, organosilane functionalization, ethylene glycol functionalization, PCP functionalization, citrate functionalization, lipoic acid functionalization, or any combination thereof. In some embodiments, the small molecule functionalization includes silica-functionalized particles, amine-functionalized particles, silicon alkoxide-functionalized particles, polystyrene-functionalized particles, and sugar-functionalized particles. In some embodiments, the small molecule functionalization includes amine-functionalization, sugar phosphate-functionalization, carboxylate-functionalization, silica-functionalization, organosilane-functionalization, or any combination thereof. In some embodiments, the small molecule functionalization includes silica-functionalization, ethylene glycol-functionalization, and amine-functionalization, or any combination thereof.

[0321] The particles of the present disclosure may be synthesized, or the particles of the present disclosure may be purchased from a commercial vendor. For example, particles consistent with the present disclosure can be purchased from commercial vendors, including Sigma-Aldrich, Life Technologies, Fisher Biosciences, nanoComposix, Nanopartz, Spherotech, and other commercial vendors. Suitable particles of the present disclosure can be purchased from commercial vendors and further modified, coated, or functionalized.

[0322] The present disclosure includes compositions and methods comprising two or more particles from among particles differing in at least one physicochemical property. Such compositions and methods may comprise at least two to at least 20 particles from among a plurality of particles differing in at least one physicochemical property. Such compositions and methods may comprise at least three to at least six particles from among a plurality of particles differing in at least one physicochemical property. Such compositions and methods may comprise at least four to at least eight particles from among a plurality of particles differing in at least one physicochemical property. Such compositions and methods may comprise at least four to at least 10 particles from among a plurality of particles differing in at least one physicochemical property. Such compositions and methods may comprise at least five to at least 12 particles from among a plurality of particles differing in at least one physicochemical property. Such compositions and methods may comprise at least six to at least 14 particles from among a plurality of particles differing in at least one physicochemical property. Such compositions and methods may comprise at least eight to at least 15 particles from among a plurality of particles differing in at least one physicochemical property. Such compositions and methods may comprise at least 10 to at least 20 particles from a plurality of particles differing in at least one physicochemical property. Such compositions and methods may include at least two different particle types, at least three different particle types, at least four different particle types, at least five different particle types, at least six different particle types, at least seven different particle types, at least eight different particle types, at least nine different particle types, at least ten different particle types, at least eleven different particle types, at least thirteen different particle types, at least fourteen different particle types, at least fifteen different particle types, at least twenty different particle types, at least twenty-five particle types, or at least thirty different particle types.

[0323] The particles of the present disclosure can be contacted with a biological sample (e.g., a biological fluid) to form a biomolecular corona. Upon contact with a complex biological sample, one or more types of particles may adsorb 100 or more types of proteins (e.g., approximately 10 of a given type in a 100 μl aliquot of a biological sample containing 100 pM of particles of one type). 10 (The particles can collectively adsorb over 100 types of proteins.) The particles and biomolecular corona can be separated from the biological sample by, for example, centrifugation, magnetic separation, filtration, or gravity separation. Particle types and biomolecular corona can be separated from the biological sample using multiple separation techniques. Non-limiting examples of separation techniques include magnetic separation, column-based separation, filtration, spin-column-based separation, centrifugation, ultracentrifugation, density- or gradient-based centrifugation, gravity separation, or any combination thereof. Protein corona analysis can be performed on the separated particles and biomolecular corona. Protein corona analysis may include identifying one or more proteins in the biomolecular corona, for example, by mass spectrometry. The method may include contacting a single particle type (e.g., particles of the types listed in Table 1) with the biological sample. The method may also include contacting multiple particle types (e.g., multiple particle types provided in Table 1) with the biological sample. Multiple particle types can be combined and contacted with the biological sample within a single sample volume. Multiple particle types can be contacted sequentially with a biological sample and separated from the biological sample before contacting subsequent particle types with the biological sample. Protein corona analysis of biomolecular coronas can compress the dynamic range of analysis compared to whole protein analysis methods.

[0324] Contacting the biological sample with a particle or particles may include adding a predetermined concentration of particles to the biological sample. Contacting the biological sample with a particle or particles may include adding 1 pM to 100 nM of particles to the biological sample. Contacting the biological sample with a particle or particles may include adding 1 pM to 500 pM of particles to the biological sample. Contacting the biological sample with a particle or particles may include adding 10 pM to 1 nM of particles to the biological sample. Contacting the biological sample with a particle or particles may include adding 100 pM to 10 nM of particles to the biological sample. Contacting the biological sample with a particle or particles may include adding 500 pM to 100 nM of particles to the biological sample. Contacting the biological sample with a particle or particles may include adding 50 μg / ml to 300 μg / ml (particle mass to biological sample volume) of particles to the biological sample. Contacting the biological sample with the particle or particles may include adding 100 μg / ml to 500 μg / ml of particles to the biological sample. Contacting the biological sample with the particle or particles may include adding 250 μg / ml to 750 μg / ml of particles to the biological sample. Contacting the biological sample with the particle or particles may include adding 400 μg / ml to 1 mg / ml of particles to the biological sample. Contacting the biological sample with the particle or particles may include adding 600 μg / ml to 1.5 mg / ml of particles to the biological sample. Contacting the biological sample with the particle or particles may include adding 800 μg / ml to 2 mg / ml of particles to the biological sample. Contacting the biological sample with the particle or particles may include adding 1 mg / ml to 3 mg / ml of particles to the biological sample. Contacting the biological sample with the particle or particles can include adding between 2 mg / ml and 5 mg / ml of particles to the biological sample. Contacting the biological sample with the particle or particles can include adding greater than 5 mg / ml of particles to the biological sample.

[0325] Particles within a plurality of particles may have varying degrees of size and shape uniformity. The standard deviation of diameter for a collection of particles of a particular type may be less than 20%, 10%, 5%, or 2% of the mean diameter of that particle type (e.g., less than 2 nm for particles with a mean diameter of 100 nm). This may correspond to a low polydispersity index for the sample containing the plurality of particles, i.e., less than 2, less than 1, less than 0.8, less than 0.6, less than 0.5, less than 0.4, less than 0.3, less than 0.2, less than 0.1, or less than 0.05. Conversely, the plurality of particles may have a large dispersity in average size and shape. The polydispersity index of a sample containing the plurality of particles may be greater than 3, greater than 4, greater than 5, greater than 8, greater than 10, greater than 12, greater than 15, or greater than 20. The size and shape uniformity among the plurality of particles may affect the number and types of biomolecules adsorbed to the particles. In some methods, particle-to-particle size uniformity (e.g., low polydispersity index) allows for higher enrichment of specific biomolecules and a stronger correspondence between the abundance of the enriched biomolecule and particle type. In some methods, lower size uniformity allows for the collection of a greater number of types of biomolecules.

[0326] Disclosed herein are methods that include obtaining a dataset including proteins detected within a biomolecular corona corresponding to physiochemically distinct particles incubated with a biological sample. The biological sample may include a blood sample from which red blood cells have been removed (e.g., an acellular sample). Physiochemically distinct types of particles produce different biomolecular coronas. Physiochemically distinct types of particles produce different biomarkers. Physiochemically distinct types of particles produce different mass spectral patterns.

[0327] Particles Panel The present disclosure provides compositions and methods for assaying protein samples. The compositions described herein include particle panels containing one or more different particle types. The particle panels described herein can vary in the number of particle types and the diversity of particle types within a single panel. For example, particles within a panel can vary based on size, polydispersity, shape and morphology, surface charge, surface chemistry and functionalization, and substrate. The panel can be incubated with a sample for analysis of proteins and protein concentration. Proteins in the sample adsorb to the surfaces of various particle types within the particle panel, forming protein coronas. The exact proteins and protein concentrations adsorbed to a particular particle type within the particle panel may depend on the composition, size, and surface charge of that particle type. Thus, each particle type within a panel can have a different protein corona to adsorb a different set of proteins, different concentrations of a particular protein, or a combination thereof. Each particle type within a panel can have mutually exclusive or overlapping protein coronas. The overlapping protein coronas can overlap in protein identity, protein concentration, or both.

[0328] The present disclosure also provides methods for selecting particle types to include in a panel depending on the type of sample. The particle types included in the panel can be a combination of particles optimized for removal of high abundance proteins. Compatible particle types for inclusion in the panel are also those selected for adsorption of specific proteins of interest. The particles can be nanoparticles. The particles can be microparticles. The particles can be a combination of nanoparticles and microparticles.

[0329] The particle panels disclosed herein can be used to identify distinct proteins disclosed herein and / or any number of specific proteins disclosed herein across a wide dynamic range. For example, particle panels disclosed herein comprising different particle types can enrich for proteins in a sample across the entire dynamic range over which proteins are present in the sample (e.g., a plasma sample). In some cases, particle panels comprising any number of different particle types disclosed herein enrich for proteins across a dynamic range of at least two orders of magnitude. In some cases, particle panels comprising any number of different particle types disclosed herein enrich for proteins across a dynamic range of at least three orders of magnitude. In some cases, particle panels comprising any number of different particle types disclosed herein enrich for proteins across a dynamic range of at least four orders of magnitude. In some cases, particle panels comprising any number of different particle types disclosed herein enrich for proteins across a dynamic range of at least five orders of magnitude. In some cases, particle panels comprising any number of different particle types disclosed herein enrich for proteins across a dynamic range of at least six orders of magnitude. In some cases, particle panels comprising any number of different particle types disclosed herein enrich for proteins across a dynamic range of at least seven orders of magnitude. In some cases, particle panels comprising any number of different particle types disclosed herein enrich proteins over a dynamic range of at least 8 orders of magnitude. In some cases, particle panels comprising any number of different particle types disclosed herein enrich proteins over a dynamic range of at least 9 orders of magnitude. In some cases, particle panels comprising any number of different particle types disclosed herein enrich proteins over a dynamic range of at least 10 orders of magnitude. In some cases, particle panels comprising any number of different particle types disclosed herein enrich proteins over a dynamic range of at least 11 orders of magnitude.In some cases, particle panels comprising any number of different particle types disclosed herein enrich for proteins over a dynamic range of at least 12 orders of magnitude. In some cases, particle panels comprising any number of different particle types disclosed herein enrich for proteins over a dynamic range of 3-5 orders of magnitude. In some cases, particle panels comprising any number of different particle types disclosed herein enrich for proteins over a dynamic range of 3-6 orders of magnitude. In some cases, particle panels comprising any number of different particle types disclosed herein enrich for proteins over a dynamic range of 4-8 orders of magnitude. In some cases, particle panels comprising any number of different particle types disclosed herein enrich for proteins over a dynamic range of 5-8 orders of magnitude. In some cases, particle panels comprising any number of different particle types disclosed herein enrich for proteins over a dynamic range of 6-10 orders of magnitude. In some cases, particle panels comprising any number of different particle types disclosed herein enrich for proteins over a dynamic range of 8-12 orders of magnitude. For example, a particle panel can collect proteins in a sample at mM and fM concentrations, thereby enriching proteins over a 12-order range.

[0330] A particle panel comprising any number of different particle types disclosed herein enriches for a single protein or a group of proteins. In some cases, the single protein or group of proteins may include proteins with different post-translational modifications. For example, a first particle type of the particle panel may enrich for a protein or group of proteins with a first post-translational modification, a second particle type of the particle panel may enrich for the same protein or group of proteins with a second post-translational modification, and a third particle type of the particle panel may enrich for the same protein or group of proteins lacking the post-translational modification. In some cases, a particle panel comprising any number of different particle types disclosed herein enriches for a single protein or group of proteins by binding different domains, sequences, or epitopes of the single protein or group of proteins. For example, a first particle type of the particle panel may enrich for a protein or group of proteins by binding to a first domain of the protein or group of proteins, and a second particle type of the particle panel may enrich for the same protein or group of proteins by binding to a second domain of the protein or group of proteins.

[0331] The particle panel may include a combination of particles having a silica surface and a polymer surface. For example, the particle panel may include SPIONs coated with a thin layer of silica, SPIONs coated with poly(dimethylaminopropylmethacrylamide) (PDMAPMA), and SPIONs coated with poly(ethylene glycol) (PEG). A particle panel consistent with the present disclosure may also include two or more particles selected from the group consisting of silica-coated SPIONs, N-(3-trimethoxysilylpropyl)diethylenetriamine-coated SPIONs, PDMAPMA-coated SPIONs, carboxyl-functionalized polyacrylic acid-coated SPIONs, amino-surface-functionalized SPIONs, polystyrene carboxyl-functionalized SPIONs, silica particles, and dextran-coated SPIONs. Particle panels consistent with the present disclosure may also include two or more particles selected from the group consisting of surfactant-free carboxylate microparticles, carboxyl-functionalized polystyrene particles, silica-coated particles, silica particles, dextran-coated particles, oleic acid-coated particles, boronated nanopowder-coated particles, PDMAPMA-coated particles, poly(glycidyl methacrylate-benzylamine)-coated particles, and poly(N-[3-(dimethylamino)propyl]methacrylamide-co-[2-(methacryloyloxy)ethyl]dimethyl-(3-sulfopropyl)ammonium hydroxide, P(DMAPMA-co-SBMA)-coated particles. Particle panels consistent with the present disclosure may also include silica-coated particles, N-(3-trimethoxysilylpropyl)diethylenetriamine-coated particles, poly(N-(3-(dimethylamino)propyl)methacrylamide) (PDMAPMA)-coated particles, sugar-phosphate-functionalized polystyrene particles, amine-functionalized polystyrene particles, polystyrene carboxyl-functionalized particles, ubiquitin-functionalized polystyrene particles, dextran-coated particles, or any combination thereof.

[0332] The particle panels disclosed herein can be used to identify multiple proteins, peptides, or protein groups using the workflow described herein (MS analysis of individual biomolecular coronas corresponding to individual particle types within the particle panel, collectively referred to as the "Proteograph" workflow). The feature intensities disclosed herein are derived from the intensities of discrete spikes ("features") found in a mass-to-charge ratio vs. intensity plot from a mass spectrometry run of a sample. These features can correspond to variably ionized fragments of peptides and / or proteins. Using the data analysis methods described herein, feature intensities can be classified into protein groups. A protein group refers to two or more proteins identified by a shared peptide sequence. Alternatively, a protein group can refer to a single protein identified using a unique identifying sequence. For example, if a peptide sequence shared between two proteins (protein 1: XYZZX and protein 2: XYZYZ) in a sample is assayed, the protein group could be the "XYZ protein group" with two members (protein 1 and protein 2). Alternatively, if the peptide sequence is unique to a single protein (protein 1), the protein group may be a "ZZX" protein group with one member (protein 1). Each protein group may be supported by multiple peptide sequences. A protein detected or identified according to the present disclosure can refer to a distinct protein detected in a sample (e.g., a distinct relative protein detected using mass spectrometry). Thus, analyzing proteins present in distinct coronas corresponding to distinct particle types within a particle panel will yield a large number of feature intensities. This number decreases as feature intensities are processed into distinct peptides, further decreases as distinct peptides are processed into distinct proteins, and further decreases when peptides are grouped into protein groups (two or more proteins that share a distinct peptide sequence).

[0333] Particle panels disclosed herein for assessing the presence or absence of one or more biomarkers associated with lung cancer (e.g., NSCLC) include at least 1 different particle type, at least 2 different particle types, at least 3 different particle types, at least 4 different particle types, at least 5 different particle types, at least 6 different particle types, at least 7 different particle types, at least 8 different particle types, at least 9 different particle types, at least 10 different particle types, at least 11 different particle types, at least 12 different particle types, at least 13 different particle types, at least 14 different particle types, at least 15 different particle types, at least 16 different particle types, at least 17 different particle types, at least 18 different particle types, at least 19 different particle types, at least 20 different particle types, at least 25 different particle types, at least 30 different particle types, at least 35 different particle types, at least 40 different particle types, at least 45 different particle types, at least 50 ... different particle types, at least 55 different particle types, at least 60 different particle types, at least 65 different particle types, at least 70 different particle types, at least 75 different particle types, at least 80 different particle types, at least 85 different particle types, at least 90 different particle types, at least 95 different particle types, at least 100 different particle types, 1-5 different particle types, 5-10 different particle types, 10-15 different particle types, 15-20 different particle types, 20-25 different particle types particle types, 25-30 different particle types, 30-35 different particle types, 35-40 different particle types, 40-45 different particle types, 45-50 different particle types, 50-55 different particle types, 55-60 different particle types, 60-65 different particle types, 65-70 different particle types, 70-75 different particle types, 75-80 different particle types, 80-85 different particle types, 85-90 different particle types, 90-95 different particle types, 95-100 different particle types, 1-100 different particle types,The number of different particle types may be 20 to 40, 5 to 10, 3 to 7, 2 to 10, 6 to 15, or 10 to 20. In certain embodiments, the present disclosure provides a panel size with 3 to 10 different particle types. In certain embodiments, the present disclosure provides a panel size with 4 to 11 different particle types. In certain embodiments, the present disclosure provides a panel size with 5 to 15 different particle types. In certain embodiments, the present disclosure provides a panel size with 5 to 15 different particle types. In certain embodiments, the present disclosure provides a panel size with 8 to 12 different particle types. In certain embodiments, the present disclosure provides a panel size with 9 to 13 different particle types. In certain embodiments, the present disclosure provides a panel size with 10 different particle types. The particle types may include nanoparticle types.

[0334] Particle panels can be designed to broadly profile proteomes, such as the human plasma proteome. A major challenge in analyzing the human proteome is that more than 99% of the approximately 3,500 protein masses in human plasma are accounted for by only 20 proteins. Plasma analysis methods are often saturated with these 20 proteins, minimizing the depth of profiling for the remaining proteins. The particle panels of the present disclosure may include particle combinations that facilitate the collection of at least 200, at least 300, at least 400, at least 500, at least 600, at least 700, at least 800, at least 900, at least 1000, at least 1100, at least 1200, at least 1300, at least 1400, at least 1500, at least 1600, at least 1700, at least 1800, at least 1900, at least 2000, at least 2100, or at least 2200 different proteins from a single biological sample. The particle panels of the present disclosure may include combinations of particles that facilitate collection of at least 4%, at least 5%, at least 6%, at least 8%, at least 10%, at least 12%, at least 15%, at least 20%, at least 25%, at least 30%, at least 35%, at least 40%, at least 45%, at least 50%, at least 55%, at least 60%, at least 65%, or at least 70% of protein types from a complex biological sample, such as human plasma. This can be achieved by providing multiple particles (e.g., as a particle panel) with different protein binding profiles. A particle panel may include two particles that, upon contact with a biological sample, form a protein corona with less than 80%, 70%, 60%, 50%, 40%, 30%, 25%, 20%, 15%, or 10% of the proteins in common. In some cases, the biological sample is human plasma.

[0335] Increasing the number of particle types in the panel increases the number of proteins that can be identified in a given sample. An example of how increasing the panel size can increase the number of identified proteins is shown in Figure 53, where a panel size of 1 particle type identified 419 different proteins, a panel size of 2 particle types identified 588 different proteins, a panel size of 3 particle types identified 727 different proteins, a panel size of 4 particle types identified 844 proteins, a panel size of 5 particle types identified 934 different proteins, a panel size of 6 particle types identified 1008 different proteins, a panel size of 7 particle types identified 1075 different proteins, a panel size of 8 particle types identified 1133 different proteins, a panel size of 9 particle types identified 1184 different proteins, a panel size of 10 particle types identified 1230 different proteins, a panel size of 11 particle types identified 1275 different proteins, and a panel size of 12 particle types identified 1318 different proteins.

[0336] Dynamic Range Some methods described herein (e.g., biomolecular corona analysis) may involve assaying biomolecules in a sample of the present disclosure over a wide dynamic range. The dynamic range of a biomolecule assayed in a sample may be the range of biomolecule abundance measured by an assay method (e.g., mass spectrometry, chromatography, gel electrophoresis, spectroscopy, or immunoassay) for the biomolecule contained in the sample. For example, an assay capable of detecting proteins over a wide dynamic range may be able to detect proteins with very low abundance to very high abundance. The dynamic range of an assay may be directly related to the slope of the assay signal intensity as a function of biomolecule abundance. For example, an assay with a low dynamic range may have a low (but positive) slope of the assay signal intensity as a function of biomolecule abundance; e.g., the ratio of the signal detected for a high-abundance biomolecule to the signal detected for a low-abundance biomolecule may be lower in an assay with a low dynamic range than in an assay with a high dynamic range. In certain cases, dynamic range may refer to the dynamic range of proteins within a sample or assay method.

[0337] The methods described herein may compress the dynamic range of an assay. The dynamic range of one assay may be compressed relative to another if the slope of the assay signal intensity as a function of biomolecule abundance is lower than that of another assay. For example, a plasma sample assayed using protein corona analysis in conjunction with mass spectrometry may have a compressed dynamic range compared to a plasma sample assayed using mass spectrometry alone, directly on the sample, or compared to the abundance values ​​of plasma proteins provided in a database (e.g., the database provided in Keshishiane et al., Mol. Cell Proteomics 14, 2375-2393 (2015), also referred to herein as the "Carr database"). The compressed dynamic range may allow for the detection of more low-abundance biomolecules in a biological sample using biomolecular corona analysis in conjunction with mass spectrometry than using mass spectrometry alone.

[0338] Collecting biomolecules on particles prior to analysis (e.g., mass spectrometry or ELISA analysis) can compress the dynamic range of the analysis. 6 The two proteins, present in a ratio of 1:1, are differentially adsorbed onto the particles, resulting in a new ratio of 10 4 The adsorption is then eluted into the solution at a concentration of 1:1. Such differential adsorption may enable the simultaneous detection of two biomolecules with a concentration difference that exceeds the dynamic range of the analytical technique. For example, mass spectrometry is often limited to measuring species within a concentration range of 4-6 orders of magnitude, making it difficult to measure species within a 10 8Two biomolecules present at a 2-fold concentration difference may not be simultaneously detected. Biomolecular corona-based enrichment of a sample can concentrate a biomolecule (e.g., a first protein) that is dilute relative to a second biomolecule (e.g., a second protein), thereby enabling simultaneous detection of two biomolecules using a single analytical method. Similarly, particle-based enrichment may enable quantification of biomolecules present at low concentrations in a sample. The dynamic range over which an analyte can be quantified is often narrower than the dynamic range over which the analyte can be detected. For example, ELISA often provides accurate concentration quantification over less than two orders of magnitude, while covering a dynamic range spanning two to three orders of magnitude. Particle-based enrichment can increase the number of biomolecular targets within a desired concentration range, thereby enabling simultaneous quantification of two or more biomolecules present in a biological sample at concentrations outside the dynamic range of the analytical technique.

[0339] Thus, various methods of the present disclosure involve detecting two biomolecules present in a biological sample at concentrations that differ by a greater amount than the dynamic range of the detection method. Many of the biomarker pairs disclosed herein span a concentration range that exceeds the detection limits of biomolecular analytical techniques (e.g., immunostaining or LC-MS / MS) and therefore may be unidentifiable or quantifiable without the use of the enrichment-based methods of the present disclosure. In some cases, the methods of the present disclosure involve detecting two biomolecules (e.g., two proteins) at concentrations that differ by at least three orders of magnitude in a biological sample (e.g., 1 mg / ml and 1 μg / ml, or 50 μM and 50 nM). In some cases, the methods of the present disclosure involve detecting two biomolecules (e.g., two proteins) at concentrations that differ by at least four orders of magnitude in a biological sample (e.g., 1 mg / ml and 100 ng / ml, or 50 μM and 5 nM). In some cases, the methods of the present disclosure involve detecting two biomolecules (e.g., two proteins) at concentrations that differ by at least five orders of magnitude in a biological sample (e.g., detecting HBA and NOTUM in human plasma). In some cases, the disclosed methods include detecting two biomolecules (e.g., two proteins) in a biological sample at concentrations that differ by at least five orders of magnitude (e.g., detecting ITIH2 and ANGL6 in human plasma). In some cases, the disclosed methods include detecting two biomolecules (e.g., two proteins) in a biological sample at concentrations that differ by at least six orders of magnitude (e.g., detecting HBA and NOTUM in human plasma). In some cases, the disclosed methods include detecting two biomolecules (e.g., two proteins) in a biological sample at concentrations that differ by at least seven orders of magnitude (e.g., detecting ceruloplasmin and RLA2 in human plasma). In some cases, the disclosed methods include detecting two biomolecules (e.g., two proteins) in a biological sample at concentrations that differ by at least seven orders of magnitude (e.g., detecting human serum albumin and CAN2 in human plasma).In some cases, the methods of the present disclosure include detecting two biomolecules (e.g., two proteins) in a biological sample at concentrations that differ by at least seven orders of magnitude (e.g., detecting human serum albumin and interleukin-6 in human plasma).

[0340] The dynamic range of a proteomic analysis assay can be the ratio of the signal generated by the most abundant protein (e.g., the protein with the highest abundance of 10%) to the signal generated by the least abundant protein (e.g., the protein with the lowest abundance of 10%). Compressing the dynamic range of a proteomic analysis can include reducing the ratio of the signal generated by the most abundant protein to the signal generated by the least abundant protein for a first proteomic analysis assay compared to that for a second proteomic analysis assay. The protein corona analysis assays disclosed herein can compress the dynamic range compared to the dynamic range of all protein analysis methods (e.g., mass spectrometry, gel electrophoresis, or liquid chromatography).

[0341] Provided herein are several methods for compressing the dynamic range of a biomolecular analytical assay to facilitate the detection of low-abundance biomolecules relative to high-abundance biomolecules. For example, the particle types disclosed herein can be used to continuously interrogate a sample. When the particle types are incubated in the sample, a biomolecular corona forms on the surface of the particle types. When biomolecules in a sample are directly detected without the use of the particle types, for example, by direct mass spectrometry of the sample, the dynamic range may span a wider concentration range, or more orders of magnitude, than when biomolecules are directly detected on the surface of the particle types. Thus, the use of the particle types disclosed herein can be used to compress the dynamic range of biomolecules in a sample. Without being limited by theory, this effect may be observed due to the particle types' high affinity in the biomolecular corona, resulting in greater capture of low-abundance biomolecules, and the particle types' low affinity in the biomolecular corona, resulting in less capture of high-abundance biomolecules.

[0342] The dynamic range of a proteomic analysis assay can be the slope of a plot of the protein signal measured by a proteomic analysis assay as a function of the total abundance of the protein in the sample. Compressing the dynamic range can include reducing the slope of a plot of the protein signal measured by a proteomic analysis assay as a function of the total abundance of the protein in the sample relative to the slope of a plot of the protein signal measured by a second proteomic analysis assay as a function of the total abundance of the protein in the sample. The protein corona analysis assay disclosed herein can compress the dynamic range compared to the dynamic range of all protein analysis methods (e.g., mass spectrometry, gel electrophoresis, or liquid chromatography).

[0343] Biomarker analysis in biological samples The methods disclosed herein can identify multiple biomarkers in biological samples (e.g., biological fluids). Non-limiting examples of biological samples that can be analyzed using the methods described herein (e.g., protein corona analysis) include biological fluid samples (e.g., fluids from cerebrospinal fluid (CSF), synovial fluid (SF), urine, plasma, serum, tears, semen, whole blood, milk, nipple aspirate, ductal lavage, vaginal fluid, nasal secretion, ear fluid, gastric juice, pancreatic juice, trabecular fluid, lung lavage, prostatic fluid, sputum, feces, bronchial lavage, swab, bronchial aspirate, sweat, or saliva), fluid solids (e.g., tissue homogenates), or cell culture-derived samples. For example, the particles disclosed herein can be incubated with any of the biological samples disclosed herein to produce at least 100 unique proteins, at least 120 unique proteins, at least 140 unique proteins, at least 160 unique proteins, at least 180 unique proteins, at least 200 unique proteins, at least 220 unique proteins, at least 240 unique proteins, at least 260 unique proteins, at least 280 unique proteins, at least 300 unique proteins, at least 320 unique proteins, at least 340 unique proteins, at least 360 unique proteins, at least 380 unique proteins, at least 400 unique proteins, at least 420 unique proteins, at least At least 440 unique proteins, at least 460 unique proteins, at least 480 unique proteins, at least 500 unique proteins, at least 520 unique proteins, at least 540 unique proteins, at least 560 unique proteins, at least 580 unique proteins, at least 600 unique proteins, at least 620 unique proteins, at least 640 unique proteins, at least 660 unique proteins, at least 680 unique proteins, at least 700 unique proteins, at least 720 unique proteins, at least 740 unique proteins, at least 760 unique proteins, at least 780 unique proteins, at least 800 unique proteins, at least 820 unique proteins,at least 840 unique proteins, at least 860 unique proteins, at least 880 unique proteins, at least 900 unique proteins, at least 920 unique proteins, at least 940 unique proteins, at least 960 unique proteins, at least 980 unique proteins, at least 1000 unique proteins, 100-1000 unique proteins, 150-950 unique proteins, 200-900 unique proteins, 250-850 unique proteins, 300-800 unique proteins, 350-750 unique proteins, 400-700 unique proteins, 450-650 unique proteins, 500-60 Protein coronas can be formed containing 0 unique proteins, 200-250 unique proteins, 250-300 unique proteins, 300-350 unique proteins, 350-400 unique proteins, 400-450 unique proteins, 450-500 unique proteins, 500-550 unique proteins, 550-600 unique proteins, 600-650 unique proteins, 650-700 unique proteins, 700-750 unique proteins, 750-800 unique proteins, 800-850 unique proteins, 850-900 unique proteins, 900-950 unique proteins, and 950-1000 unique proteins. In some cases, a similar number of proteins can be assessed without the use of particles or using the assay methods described herein. In some embodiments, several different types of particles can be used individually or in combination to identify a large number of proteins in a particular biological sample. In other words, the particles can be multiplexed to bind and identify multiple proteins in a biological sample.

[0344] The methods disclosed herein can be used to identify various biological states in a particular biological sample. For example, a biological state may refer to elevated or low levels of a particular protein or set of proteins, or may be evidenced by the ratio between the abundances of two or more biological molecules. In another example, a biological state may refer to the identification of a disease, such as cancer. A biological state may include cancerous lung nodules. A biological state may include non-cancerous lung nodules. Incubating one or more particle types with a biological sample, such as human plasma, allows for the formation of a protein corona. The protein corona can then be analyzed to identify a protein pattern. Analysis may include gel electrophoresis, mass spectrometry, chromatography, ELISA, immunohistology, or any combination thereof. Analysis of the protein corona (e.g., by mass spectrometry or gel electrophoresis) may be referred to as corona analysis. The protein pattern can be compared with the same method performed on a control sample. Comparing the protein patterns can identify that a first sample contains elevated levels of markers corresponding to a particular type of lung cancer. Thus, the particles and methods of use thereof can be used to diagnose specific disease conditions.

[0345] The assay may include protein collection of the particles, protein digestion, and mass spectrometry (e.g., MS, LC-MS, LC-MS / MS). Digestion may include chemical digestion, such as with cyanogen bromide or 2-nitro-5-thiocyanatobenzoic acid (NTCB). Digestion may include enzymatic digestion, such as with trypsin or pepsin. Digestion may include enzymatic digestion with multiple proteases. Digestion may involve a protease selected from the group consisting of trypsin, chymotrypsin, GluC, LysC, elastase, subtilisin, proteinase K, thrombin, factor X, ArgC, papain, AspN, thermolysin, pepsin, aspartyl protease, cathepsin D, zinc mealyloprotease, glycoprotein endopeptidase, proline, aminopeptidase, prenyl protease, caspase, kex2 endoprotease, or any combination thereof. The digestion method may cleave peptides randomly or at a specific position or series of positions. Multiple digestion methods (e.g., two or more proteases) can be utilized in an assay. The assay may involve dividing a sample into multiple portions and subjecting the portions to different digestion methods and separate analyses (e.g., separate mass spectrometry analyses). Digestion can cleave peptides at specific positions (e.g., methionine) or sequences (e.g., glutamic acid-histidine-glutamic acid). Digestion can sometimes distinguish between similar proteins. For example, in an assay, a first digestion method can resolve eight distinct proteins as a single protein group, while a second digestion method can resolve eight distinct proteins with distinct signals. Digestion can generate peptide fragments with an average length of 8-15 amino acids. Digestion can generate peptide fragments with an average length of 12-18 amino acids. Digestion can generate peptide fragments with an average length of 15-25 amino acids. Digestion can generate peptide fragments with an average length of 20-30 amino acids. Digestion can generate peptide fragments with an average length of 30-50 amino acids.

[0346] Various methods of the present disclosure enable measurements across a wide concentration range. Analytical methods for biomolecules are often limited to a narrow concentration range. For example, proteomic analysis by mass spectrometry is often limited to three, four, or five orders of magnitude. Therefore, the presence of a biomolecule at a relatively high concentration (e.g., at mg / ml concentrations) may mask the detection of a biomolecule at a low concentration, further limiting the accuracy of quantification of the low-concentration biomolecule. The methods of the present disclosure may enable the detection of molecules whose concentrations span at least five, at least six, at least seven, at least eight, at least nine, at least ten, at least eleven, or at least twelve orders of magnitude. Thus, the methods of the present disclosure can detect and quantify both biomolecules at relatively high and relatively low concentrations from a single sample without first depleting the biomolecule from the sample. For example, a plasma assay consistent with the present disclosure can simultaneously quantify albumin (present at approximately 40 mg / ml) and interleukin-10 (present at approximately 6 pg / ml) from a single undepleted plasma sample, thereby simultaneously detecting two species that differ in concentration by approximately 10 orders of magnitude.

[0347] Biomarkers for cancer detection Proteins may be included as biomarkers for disease detection. Disease detection may include cancer detection through the use of biomarkers such as proteins. Proteins may be generated as part of protein data or proteomic data.

[0348] Example proteins can include any of the proteins in Figures 26A-26B. Protein data may include measurements of 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, or 15 of these proteins, or measurements in any range of the aforementioned number of proteins from these figures.

[0349] Some examples of proteins are shown in Figure 30A. Proteins that can be detected using the methods described herein include myosin-9 (MYH9), tubulin beta-1 chain (TUBB1), tubulin beta chain (TUBB), calreticulin (CALR), vascular endothelial growth factor receptor 3 (FLT4), neurogenic locus notch homolog protein 2 (NOTCH2), transforming protein RhoA (RHOA), isocitrate dehydrogenase [NADP], mitochondrial (IDH2), cadherin 1 (CDH1), cAMP-dependent protein 1 (CAMP1), and phosphodiesterase (CP1). Examples of proteins include protease kinase type I-alpha regulatory subunit (PRKAR1A), neurogenic locus notch homolog protein 1 (NOTCH1), exostosin-1 (EXT1), serine / threonine protein phosphatase 2A 65 kDa regulatory subunit A alpha isoform (PPP2R1A), staphylococcal nuclease domain-containing protein 1 (SND1), tyrosine protein kinase BTK (BTK), lipoma-preferred partner (LPP), mitogen-activated protein kinase (MAPK1), Fat1 protein (FAT1), cadherin-11 (CDH11), or dual specificity mitogen-activated protein kinase kinase 1 (MAP2K1). Additional examples of proteins are shown in Figures 32A-32B. Proteins detected by the methods described herein may also include thrombospondin-2 (TSP2 or P35442). Further examples of proteins are shown in Figures 32C-32D. Proteins detected by the methods described herein may include P01011. Some examples of proteins are shown in Figure 36. Proteins detected by the methods described herein include polymeric immunoglobulin receptor (PIGR, UniProt P01833), cadherin-related family member 2 (CDHR2, UniProt Q9BYE9), leucine-rich alpha-2 glycoprotein (LRG1 or A2GL, UniProt P02750), intercellular adhesion molecule 1 (ICAM1, UniProt P05362), aminopeptidase N (AMPN or ANPEP, UniProtP15144), thrombospondin-2 (TSP2, UniProt P35442), protein S100-A9 (S10A9 or S100A9, UniProt P06702), aldo-keto reductase family 1 member B1 (ALDR or AKR1B1, UniProt P15121), serum amyloid A-1 protein (SAA1, UniProt P0DJI8), peroxidasin homolog (PXDN, UniProt Q92626), protein S100-A8 (S10A8 or S100A8, UniProt P05109), anthrax toxin receptor 2 (ANTR2 or ANTXR2, UniProt P58335), cadherin-2 (CADH2 or CDH2, UniProt P15121), P19022), alpha-1-antichymotrypsin (AACT or SERPINA3, UniProt P01011), collagen alpha-1(XVIII) chain (COIA1 or COL18A1, UniProt P39060), fibrinogen-like protein 1 (FGL1, UniProt Q08830), protein S100-A12 (S10AC or S100A12, UniProt P80511), reelin (RELN, UniProt J3KQ66), C-reactive protein (CRP, UniProt P02741), versican core protein (CSPG2 or VCAN, UniProt P13611), coagulation factor XIIIA chain (F13A or F13A1, UniProt P00488), cartilage intermediate lamina protein 2 (CILP2, UniProt K7EPJ4), Sushi, von Willebrand factor type A, EGF and pentraxin domain-containing protein 1 (SVEP1, UniProt Q4LDE5), neutrophil gelatinase-associated lipocalin (NGAL or LCN2, UniProt P80188), tetranectin (TETN or CLEC3B, UniProt P05452), SLAIN motif-containing protein 2 (SLAI2 or SLAIN2, UniProt Q9P270), anthrax toxin receptor 1 (ANTR1 or ANTXR1, UniProt Q9H6X2, e.g., isoform 5 [UniProt Q9H6X2-5]), or serum amyloid A-2 protein (SAA2, UniProt Q9H6X2-5]).P0DJI9). Any number of the aforementioned proteins can be used. Any of the proteins can be used in the classifier.

[0350] Exemplary proteins may include SERPINA1, HPR, EPS15L1, ORM2, CTSH, CRP, SAA4, COLEC10, HIST1H4I, APOM, ORM1, PODOX8, IGKV1-8, IGKV1-9, ANGPTL6, SERPINA3, PXDN, IGKC, HP, APCS, or ITIH2. The protein data may include measurements of 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, or 21 of these proteins, or ranges of measurements for any of the aforementioned numbers of these proteins.

[0351] The method may include measuring a biomarker in a biological fluid sample. The method may include using a biomarker in a biological fluid sample. The biomarker may include A2GL, AKR1B1, ANPEP, ANTXR1, ANTXR2, BTK, CALR, CDH1, CDH11, CDH2, CDHR2, CILP2, CLEC3B, COL18A1, CRP, EXT1, F13A1, FAT1, FGL1, FLT4, ICAM1, IDH2, LCN2, LPP, MAPK1, MAP2K1, MYH9, NOTCH1, NOTCH2, PIGR, PPP2R1A, PRKAR1A, PXDN, RELN, RHOA, S100A8, S100A9, S100A12, SAA1, SAA2, SERPINA3, SLAIN2, SND1, SVEP1, TSP2, TUBB, TUBB1, or VCAN. In some embodiments, the biomarkers include 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 20, 25, 30, 35, 40, 45, or 48 of the foregoing biomarkers, or a range defined by any two of the foregoing integers.

[0352] Proteomic data can include protein measurement values.The protein measurement values ​​in the sample from the subject with liver cancer can be increased or decreased compared with the protein measurement values ​​from control samples or compared with baseline measurement values.Protein measurement values ​​can include the measurement values ​​of the protein or protein combinations from Figure 39C or Figure 39D. For example, protein measurements can include measurements of one or more of the following proteins: 3-ketoacyl-CoA thiolase, peroxisomal (ACAA1), adenosine deaminase 2 (ADA2), angiotensinogen (AGT), acidic leucine-rich nuclear phosphoprotein 32 family member A (ANP32A), aquaporin-1 (AQP1), actin-associated protein 2 / 3 complex subunit 1B (ARPC1B), asialoglycoprotein receptor 2 (ASGR2), aspartyl / asparaginyl beta-hydroxylase (ASPH), calreticulin (CALR), F-actin capping protein subunit alpha-1 (CAPZA1), carbonyl reductase [NADPH]1 (CBR1). , CD5 antigen-like (CD5L), cell migration induction and hyaluronan binding protein (CEMIP), chordin-like protein 1 (CHRDL1), beta-Ala-His dipeptidase (CNDP1), collagen alpha-1(XIV) chain (COL14A1), collagen alpha-1(VI) chain (COL6A1), dnaJ homolog subfamily B member 11 (DNAJB11), desmocollin-2 (DSC2), desmoglein-2 (DSG2), bifunctional glutamate / proline-tRNA ligase (EPRS1), endothelial cell-specific molecule 1 (ESM1), electron transfer flavoprotein subunit beta (ETFB), fibroleukin (FGL2), four-and-a-half LIM domain protein 1 (four-and-a-half LIM domains protein 1) (FHL1), fibromodulin (FMOD), fructosamine-3-kinase (FN3K), glypican-1 (GPC1), phosphatidylinositol-glycan-specific phospholipase D (GPLD1), glyoxylate reductase / hydroxypyruvate reductase (GRHPR),Trifunctional enzyme subunit alpha, mitochondrial (HADHA), hepatoma-derived growth factor (HDGF), HLA class I histocompatibility antigen, C alpha chain (HLA.C), insulin-like growth factor binding protein complex acid-labile subunit (IGFALS), insulin-like growth factor binding protein 2 (IGFBP2), insulin-like growth factor binding protein 5 (IGFBP5), interleukin enhancer-binding factor 2 (ILF2), integrin alpha-M (ITGAM), galectin-3-binding protein (LGALS3BP), amine oxidase [flavin-containing] B (MAOB), methyltransferase-like protein 7A (METTL7A), myeloperoxidase (MPO), nicotinamide phosphoribosyltransferase (NAMPT), NIF3-like protein 1 (NIF3L1), neuropilin-1 (NRP1), nucleobindin-1 (NUCB1), beta-parvin (PARVB), profilin-1 (PFN1), glycerol-3-phosphate phosphatase (PGP), peptidase inhibitor 16 (PI16), polymeric immunoglobulin receptor (PIGR), phosphomevalonate kinase (PMVK), proteoglycan 4 (PRG4), trypsin-2 (PRSS2), 26S proteasome regulatory subunit 6B (PSMC4), pentraxin-related protein PTX3 (PTX3), peroxidasin homolog (PXDN), rab GTPase-activating protein 1 (RABGAP1), 60S ribosomal protein L12 (RPL12), 40S ribosomal protein S7 (RPS7), protein S100-A8 (S100A8), protein S100-A9 (S100A9), serum amyloid A-1 protein (SAA1), sushi, von Willebrand factor type A, EGF and pentraxin domain-containing protein 1 (SVEP1), transgelin-2 (TAGLN2), transferrin receptor protein 1 (TFRC), transforming growth factor beta-induced protein ig-h3 (TGFBI), talin-1 (TLN1), tenascin (TNC), tropomyosin alpha-1 chain (TPM1), tubulin alpha-1C chain (TUBA1C),or versican core protein (VCAN). In some embodiments, the protein is ACAA1, ADA2, AGT, ANP32A, AQP1, ARPC1B, ASGR2, ASPH, CALR, CAPZA1, CBR1, CD5L, CEMIP, CHRDL1, CNDP1, COL14A1, COL6A1, DNAJB11, DSC2, DSG2, EPRS1, ESM1, ETFB, FGL2, FHL1, FMOD, FN3K, GPC1, GPLD1, GRHPR, HADHA, HDGF, HLA.C, IGFALS, IGFBP2, IGFBP5 , ILF2, ITGAM, LGALS3BP, MAOB, METTL7A, MPO, NAMPT, NIF3L1, NRP1, NUCB1, PARVB, PFN1, PGP, PI16, PIGR, PMVK, PRG4, PRSS2, PSMC4, PTX3, PXDN, RABGAP1, RPL12, RPS7, S100A8, S100A9, SAA1, SVEP1, TAGLN2, TFRC, TGFBI, TLN1, TNC, TPM1, TUBA1C, or VCAN, or a combination thereof. A combination of proteins may include 2, 3, 4, 5, 6, 7, 8, 9, 10, 15, 20, 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, or 72 proteins in Figure 39C, or a range of proteins defined by any two of the foregoing integers. A combination of proteins may include at least 1, at least 2, at least 3, at least 4, at least 5, at least 6, at least 7, at least 8, at least 9, at least 10, at least 15, at least 20, at least 25, at least 30, at least 35, at least 40, at least 45, at least 50, at least 55, at least 60, at least 65, or at least 70 proteins in Figure 39C. A combination of proteins may include fewer than 3, fewer than 4, fewer than 5, fewer than 6, fewer than 7, fewer than 8, fewer than 9, fewer than 10, fewer than 15, fewer than 20, fewer than 25, fewer than 30, fewer than 35, fewer than 40, fewer than 45, fewer than 50, fewer than 55, fewer than 60, fewer than 65, fewer than 70, or fewer than 72 proteins in Figure 39C. In some embodiments, a combination of proteins may include:The protein does not include one or more of the proteins in Figure 39C or Figure 39D. In some embodiments, the protein includes a protein useful for assessing lung nodules, such as APP, IGHG2, SERPING1, SAA2, SERPINF2, GC, IGHA1, HPR, SERPINA3, IGHA1, LTF, SERPINA1, PCSK6, PROS1, BPIF1, C6, CP, A2M, or IGFBP2.

[0353] The proteomic data may include protein measurements. Protein measurements in samples from subjects with ovarian cancer may be increased or decreased compared to protein measurements from control samples or compared to baseline measurements. Protein measurements may include measurements of a protein or combination of proteins from Figure 40C. For example, protein measurements may include measurements of one or more of the following proteins: anthrax toxin receptor 2 (ANTXR2), bone morphogenetic protein 1 (BMP1), cartilage intermediate lamina protein 1 (CILP), interferon-induced double-stranded RNA-activated protein kinase (EIF2AK2), beta-enolase (ENO3), coagulation factor XIIIB chain (F13B), fibrinogen-like protein 1 (FGL1), or phosphatidylethanolamine-binding protein 4 (PEBP4). The protein may include ANTXR2. The protein may include BMP1. The protein may include CILP. The protein may include EIF2AK2. The protein may include ENO3. The protein may include F13B. The protein may include FGL1. The protein may include PEBP4. The protein combination may include 2, 3, 4, 5, 6, 7, or 8 proteins of FIG. 40C, or a range of proteins defined by any two of the foregoing integers. The protein combination may include at least one, at least two, at least three, at least four, at least five, at least six, or at least seven proteins of FIG. 40C. The protein combination may include fewer than three, fewer than four, fewer than five, fewer than six, fewer than seven, or fewer than eight proteins of FIG. 40C. In some embodiments, the protein combination does not include one or more of the proteins of FIG. 40C or FIG. 39E.

[0354] Described herein are biomarkers that can be analyzed by the methods described herein to determine whether a subject has a pulmonary nodule, a benign pulmonary nodule, or a malignant pulmonary nodule. In some embodiments, the biomarker is a protein. In some embodiments, the biomarker is a nucleic acid encoding any one of the proteins or peptide fragments of the proteins described herein. In some aspects, the biomarker comprises a protein, such as a secreted protein.

[0355] Biomarkers disclosed herein (e.g., associated with a disease state such as NSCLC, a comorbidity, or a healthy state) may include at least one of the following: protein S100-A9 (P06702; S10A9_HUMAN), C-reactive protein (P02741; CRP_HUMAN), inter-alpha trypsin inhibitor heavy chain H2 (P19823; ITIH2_HUMAN), protein S100-A8 (P05109; S10A8_HUMAN), serine protease HTRA1 (Q92743; HTRA1_HUMAN), and alpha trypsin inhibitor heavy chain H2 (P19823; ITIH2_HUMAN). Angiopoietin-related protein 6 (Q8NI99; ANGL6_HUMAN), haptoglobin-related protein (P00739; HPTR_HUMAN), CC motif chemokine 18 (P55774; CCL18_HUMAN), actin, cytoplasmic 1 (P60709; ACTB_HUMAN), actin, cytoplasmic 2 (P63261; ACTG_HUMAN), serum amyloid A-1 protein (P0DJI8; SAA1_HUMAN), immunoglobulin kappa constant (P01834; IGKC_HUMAN), angiopoietin-related protein 6 (Q8NI99; ANGL6_HUMAN), peroxidasin homolog (Q92743; PXDN_HUMAN), anthrax toxin receptor 2 (P58335; ANTR2_HUMAN), tubulin alpha-1A chain (Q71U36; TBA1A_HUMAN), syndecan-1 (P18827; SDC1_HUMAN), serum amyloid A-2 protein (P0DJI9; SAA2_HUMAN), versican core protein (P13611; CSPG2_HUMAN), anthrax toxin receptor 1 (Q9H6X2; ANTR1_HUMAN), palmi treoyl protein carboxylesterase NOTUM (Q6P988; NOTUM_HUMAN), cartilage intermediate lamina protein 1 (O75339; CILP1_HUMAN), calpain-2 catalytic subunit (P17655; CAN2_HUMAN), 60S acidic ribosomal protein P2 (P05387; RLA2_HUMAN), beta-galactoside alpha-2,6-sialyltransferase 1 (P15907; SIAT1_HUMAN), and platelet glycoprotein Ib beta chain (P13224; GP1BB_HUMAN).The biomarkers may include any biomarker or biomarkers in Figure 52. Any one or more of the above biomarkers can be used in various combinations to train a classifier to distinguish whether a subject has lung cancer (e.g., NSCLC), a co-morbid condition, or is healthy. Any one or more of the above biomarkers can be used in various combinations to train a classifier to distinguish whether a subject has cancerous or non-cancerous lung nodules. In some embodiments, a combination of at least one of the biomarkers, at least two of the biomarkers, at least three of the biomarkers, at least four of the biomarkers, at least five of the biomarkers, at least six of the biomarkers, at least seven of the biomarkers, at least eight of the biomarkers, at least nine of the biomarkers, at least ten of the biomarkers, at least fifteen of the biomarkers, at least twenty of the biomarkers, at least twenty-five of the biomarkers, or all of the biomarkers can be used to train a classifier to distinguish whether a subject has a cancerous or non-cancerous lung nodule. In some embodiments, at least one of the biomarkers, at least two of the biomarkers, at least three of the biomarkers, at least four of the biomarkers, at least five of the biomarkers, at least six of the biomarkers, at least seven of the biomarkers, at least eight of the biomarkers, at least nine of the biomarkers, at least ten of the biomarkers, at least fifteen of the biomarkers, at least twenty of the biomarkers, at least twenty-five of the biomarkers, or a combination of all of the biomarkers can be used in a diagnostic assay to determine whether a subject has a cancerous or non-cancerous lung nodule. The diagnostic assay can be performed using the trained classifier disclosed herein. Where the use of biomarkers is described, a biological molecule can be used.The biomarkers may include classifier features disclosed herein.

[0356] The present disclosure provides methods for detecting low-abundance peptides in complex biological samples. Many of the disclosed diagnostic peptides are inaccessible to conventional blood analysis methods due to the high concentrations of albumin, immunoglobulins, and other abundant blood proteins. Diagnostic peptides may be present at concentrations 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, or more orders of magnitude lower than the most abundant proteins in a blood sample and therefore cannot be detected by many conventional proteomics methods. The present disclosure provides methods for enriching low-abundance biomolecules (e.g., proteins) from complex biological samples such as plasma and for quantifying the enriched biomolecules.

[0357] Examples of lung cancer diagnostic peptides are shown in Table 2. Additional examples of diagnostic peptides for various cancers are shown in other figures or tables. The methods of the present disclosure may include assaying a sample from a subject to detect the presence, absence, or abundance of one or more peptides or peptide fragments from among the peptides listed in Table 2 or another table or figure provided herein. In some cases, the method includes identifying a ratio between the abundance of two peptides or peptide fragments from among the peptides listed in Table 2 or another table or figure provided herein. In some cases, the method includes identifying a ratio between the abundance of a peptide or peptide fragment from among the peptides listed in Table 2 or another table or figure provided herein and a separate peptide from the same biological sample. For example, the method may include identifying the ratio of the relative abundance of APOC1 and ceruloplasmin in a plasma sample from a subject with a lung nodule. In some cases, the method includes assaying the sample to detect the presence, absence, or abundance of one or more peptides or fragments of peptides from the group consisting of angiopoietin-related protein 6 (ANGL6), palmitoleoylprotein carboxylesterase NOTUM (NOTUM), cartilage intermediate lamina protein 1 (CILP1), 60S acidic ribosomal protein P2 (RLA2), and platelet glycoprotein Ib beta chain (GP1BB). In some cases, the method includes assaying the sample to detect the presence, absence, or abundance of at least 2, at least 3, at least 4, at least 5, at least 6, at least 8, at least 10, at least 12, at least 15, at least 20, at least 25, at least 30, or at least 35 peptides or fragments of peptides from the peptides listed in Table 2 or another table or figure provided herein.

[0358] The disclosed methods allow for the quantification of heterogeneous biomarkers across a wide concentration range. In some cases, lung cancer (e.g., NSCLC) is evidenced by the relative concentrations of two or more proteins from a patient sample. In some cases, the disclosed methods include identifying an abundance (e.g., concentration) ratio between at least two peptides from among the peptides listed in Table 2 or another table or figure provided herein. In some cases, the disclosed methods include identifying a ratio between the abundances of at least three peptides from among the peptides listed in Table 2 or another table or figure provided herein. In some cases, the disclosed methods include identifying a ratio between the abundances of at least four peptides from among the peptides listed in Table 2 or another table or figure provided herein. In some cases, the disclosed methods include identifying a ratio between the abundances of at least five peptides from among the peptides listed in Table 2 or another table or figure provided herein. In some cases, the disclosed methods include identifying a ratio between the abundances of at least six peptides from among the peptides listed in Table 2 or another table or figure provided herein. In some cases, the methods of the present disclosure include identifying a ratio between the abundances of at least seven peptides from among the peptides listed in Table 2 or another table or figure provided herein. In some cases, the methods of the present disclosure include identifying a ratio between the abundances of at least eight peptides from among the peptides listed in Table 2 or another table or figure provided herein. In some cases, the methods of the present disclosure include identifying a ratio between the abundances of at least nine peptides from among the peptides listed in Table 2 or another table or figure provided herein. In some cases, the methods of the present disclosure include identifying a ratio between the abundances of at least ten peptides from among the peptides listed in Table 2 or another table or figure provided herein. In some cases, the methods of the present disclosure include identifying a ratio between the abundances of at least twelve peptides from among the peptides listed in Table 2 or another table or figure provided herein.In some cases, the disclosed method includes identifying a ratio between the abundances of at least 15 peptides from among the peptides listed in Table 2 or another table or figure provided herein. In some cases, the disclosed method includes identifying a ratio between the abundances of at least 20 peptides from among the peptides listed in Table 2 or another table or figure provided herein. In some cases, the disclosed method includes identifying a ratio between the abundances of at least 25 peptides from among the peptides listed in Table 2 or another table or figure provided herein. In some cases, the sample is a blood sample (e.g., plasma).

[0359] In some cases, the method includes assaying the sample to detect the presence, absence, or abundance of at least two, at least three, at least four, or all five of ANGL6, NOTUM, CILP1, RLA2, or GP1BB. In some cases, the one or more peptides or fragments of peptides from among the peptides listed in Table 2 are selected from the group consisting of actin (e.g., beta-actin), anthrax toxin receptor 2, cartilage intermediate lamina protein 1, collectin 11, and kallistatin. In some cases, one or more peptides or peptide fragments from among the peptides listed in Table 2 are selected from the group consisting of angiopoietin-related protein 6 (ANGL6), serine protease HTRA1 (HTRA1), peroxidasin homolog (PXDN), CC motif chemokine 18 (CCL18), anthrax toxin receptor 2 (ANTR2), tubulin alpha-1A chain (TBA1A), syndecan-1 (SDC1), serum amyloid A-2 protein (SAA2), versican, and the like. core protein (CSPG2), anthrax toxin receptor 1 (ANTR1), palmitoleoyl protein carboxylesterase NOTUM (NOTUM), cartilage intermediate lamina protein 1 (CILP1), calpain-2 catalytic subunit (CAN2), 60S acidic ribosomal protein P2 (RLA2), beta-galactoside alpha-2,6-sialyltransferase 1 (SIAT1), and platelet glycoprotein Ib beta chain (GP1BB).In some cases, the one or more peptides or fragments of peptides from among the peptides listed in Table 2 are selected from the group consisting of: leucine-rich alpha-2-glycoprotein (A2GL), actin, cytoplasmic 1 (ACTB), actin, cytoplasmic 2 (ACTG), apolipoprotein CI (APOC1), apolipoprotein M (APOM), voltage-gated calcium channel subunit alpha-2 / delta-1 (CA2D1), cadherin-13 (CAD13), beta-Ala-His dipeptidase (CNDP1), ciliary nerve Further included are trophin receptor subunit alpha (CNTFR), collectin-11 (COL11), C-reactive protein (CRP), hemoglobin subunit alpha (HBA), haptoglobin-related protein (HPT), haptoglobin-related protein (HPTR), inter-alpha trypsin inhibitor heavy chain H2 (ITIH2), kallistatin (KAIN), plasma kallikrein (KLKB1), neural cell adhesion molecule 1 (NCAM1), protein S100-A8 (S10A8), protein S100-A9 (S10A9), and maintenance of chromosome structure protein 4 (SMC4). In some cases, one or more peptides or fragments of peptides from among the peptides listed in Table 2 are selected from the group consisting of A2GL, ACTB, ACTG, APOC1, APOM, CA2D1, CAD13, CNDP1, CNTFR, COL11, CRP, HBA, HPT, HPTR, ITIH2, KAIN, KLKB1, NCAM1, S10A8, S10A9, or SMC4.In some cases, the one or more peptides or fragments of peptides from among the peptides listed in Table 2 include at least 2, at least 3, at least 4, at least 5, at least 6, at least 7, at least 8, at least 9, at least 10, at least 11, at least 12, at least 13, at least 14, at least 15, at least 16, at least 17, at least 18, at least 19, or at least 20 of A2GL, ACTB, ACTG, APOC1, APOM, CA2D1, CAD13, CNDP1, CNTFR, COL11, CRP, HBA, HPT, HPTR, ITIH2, KAIN, KLKB1, NCAM1, S10A8, S10A9, or SMC4. [Table 2-1] [Table 2-2]

[0360] In some cases, the method includes detecting angiopoietin-related protein 6 (ANGL6), serine protease HTRA1 (HTRA1), peroxidasin homolog (PXDN), CC motif chemokine 18 (CCL18), anthrax toxin receptor 2 (ANTR2), tubulin alpha-1A chain (TBA1A), syndecan-1 (SDC1), serum amyloid A-2 protein (SAA2), versican core protein (CSPG2), anthrax toxin receptor 1 (ANTR1), palmitoyltransferase (PTT1), thrombin-dependent protein kinase C (THC ... The method includes detecting the presence, absence, or abundance of one or more peptides selected from the group consisting of treacleoyl protein carboxylesterase NOTUM (NOTUM), cartilage intermediate lamina protein 1 (CILP1), calpain-2 catalytic subunit (CAN2), 60S acidic ribosomal protein P2 (RLA2), beta-galactoside alpha-2,6-sialyltransferase 1 (SIAT1), and platelet glycoprotein Ib beta chain (GP1BB). In some cases, the method includes identifying a ratio between the abundances of two peptides selected from the group consisting of angiopoietin-related protein 6 (ANGL6), serine protease HTRA1 (HTRA1), peroxidasin homolog (PXDN), CC motif chemokine 18 (CCL18), anthrax toxin receptor 2 (ANTR2), tubulin alpha-1A chain (TBA1A), syndecan-1 (SDC1), serum amyloid A-2 protein (SAA2), versican core protein (CSPG2), anthrax toxin receptor 1 (ANTR1), palmitoleoyl protein carboxylesterase NOTUM (NOTUM), cartilage intermediate lamina protein 1 (CILP1), calpain-2 catalytic subunit (CAN2), 60S acidic ribosomal protein P2 (RLA2), beta-galactoside alpha-2,6-sialyltransferase 1 (SIAT1), and platelet glycoprotein Ib beta chain (GP1BB).In some cases, the method includes detecting angiopoietin-related protein 6 (ANGL6), serine protease HTRA1 (HTRA1), peroxidasin homolog (PXDN), CC motif chemokine 18 (CCL18), anthrax toxin receptor 2 (ANTR2), tubulin alpha-1A chain (TBA1A), syndecan-1 (SDC1), serum amyloid A-2 protein (SAA2), versican core protein (CSPG2), anthrax toxin receptor 1 (ANTR1), palmitoleoyl protein carboxylesterase NOTUM (NOTUM), soft tissue endothelial growth factor receptor 1 (STFR1), fibronectin receptor 1 (FOLFOX1), fibronectin receptor 1 (FOLFOX2), fibronectin receptor 1 (FOLFOX3), fibronectin receptor 1 (FOLFOX4), fibronectin receptor 1 (FOLFOX5), fibronectin receptor 1 (FOLFOX6), fibronectin receptor 1 (FOLFOX7), fibronectin receptor 1 (FOLFOX8), fibronectin receptor 1 (FOLFOX9), fibronectin receptor 1 (FOLFOX1), fibronectin receptor 1 (FOLFOX1), fibronectin receptor 1 (FOLFOX1), fibronectin receptor 1 (FOLFOX1), fibronectin receptor 1 (FOLFOX1), fibronectin receptor 1 (FOLFOX1), fibronectin receptor 1 (FOLFOX2 ... The method includes detecting the presence, absence, or abundance of at least two, at least three, at least four, at least five, at least six, at least eight, at least ten, at least twelve, or at least fifteen peptides selected from the group consisting of bone intermediate lamina protein 1 (CILP1), calpain-2 catalytic subunit (CAN2), 60S acidic ribosomal protein P2 (RLA2), beta-galactoside alpha-2,6-sialyltransferase 1 (SIAT1), and platelet glycoprotein Ib beta chain (GP1BB).

[0361] Biomarkers (e.g., proteins) may include angiopoietin-related proteins, serine proteases, peroxidasin homologs, CC motif chemokines, anthrax toxin receptors, tubulin proteins, syndecan proteins, serum amyloid A proteins, versican proteins, anthrax toxin receptor proteins, palmitoleoyl protein carboxylesterase proteins, cartilage intermediate lamina proteins, calpain proteins or subunits, 60S acidic ribosomal proteins, beta-galactoside alpha-2,6-sialyltransferase proteins, or platelet glycoproteins, or subunits or fragments of any of the foregoing proteins. Biomarkers may include angiopoietin-related proteins. Biomarkers may include serine proteases. Biomarkers may include peroxidasin homologs. Biomarkers may include CC motif chemokines. Biomarkers may include anthrax toxin receptors. Biomarkers may include tubulin proteins. Biomarkers may include syndecan proteins. The biomarker may include serum amyloid A protein. The biomarker may include versican protein. The biomarker may include anthrax toxin receptor protein. The biomarker may include palmitoleoyl protein carboxylesterase protein. The biomarker may include cartilage intermediate lamina protein. The biomarker may include calpain protein or subunit. The biomarker may include 60S acidic ribosomal protein. The biomarker may include beta-galactoside alpha-2,6-sialyltransferase protein. The biomarker may include platelet glycoprotein. The biomarker may be secreted.

[0362] Biomarkers (e.g., proteins) may include angiopoietin-related protein 6 (ANGL6), serine protease HTRA1 (HTRA1), peroxidasin homolog (PXDN), CC motif chemokine 18 (CCL18), anthrax toxin receptor 2 (ANTR2), tubulin alpha-1A chain (TBA1A), syndecan-1 (SDC1), serum amyloid A-2 protein (SAA2), versican core protein (CSPG2), anthrax toxin receptor 1 (ANTR1), palmitoleoyl protein carboxylesterase NOTUM (NOTUM), cartilage intermediate lamina protein 1 (CILP1), calpain-2 catalytic subunit (CAN2), 60S acidic ribosomal protein P2 (RLA2), beta-galactoside alpha-2,6-sialyltransferase 1 (SIAT1), or platelet glycoprotein Ib beta chain (GP1BB). Biomarkers (e.g., proteins) may include angiopoietin-related protein 6 (ANGL6), serine protease HTRA1 (HTRA1), peroxidasin homolog (PXDN), CC motif chemokine 18 (CCL18), anthrax toxin receptor 2 (ANTR2), tubulin alpha-1A chain (TBA1A), syndecan-1 (SDC1), serum amyloid A-2 protein (SAA2), versican core protein (CSPG2), anthrax toxin receptor 1 (ANTR1), palmitoleoyl protein carboxylesterase NOTUM (NOTUM), cartilage intermediate lamina protein 1 (CILP1), calpain-2 catalytic subunit (CAN2), 60S acidic ribosomal protein P2 (RLA2), beta-galactoside alpha-2,6-sialyltransferase 1 (SIAT1), and platelet glycoprotein Ib beta chain (GP1BB).

[0363] In some embodiments, the biomarker is a secreted protein. In some embodiments, the biomarker comprises a protein involved in a metabolic pathway. In some embodiments, the biomarker comprises a protein involved in oxidative phosphorylation.

[0364] In some embodiments, the biomarker comprises cell-free RNA. In some embodiments, the biomarker is RNA encoding a secreted protein. In some embodiments, the biomarker comprises mRNA encoding a protein involved in a metabolic pathway. In some embodiments, the biomarker comprises mRNA encoding a protein involved in oxidative phosphorylation.

[0365] The biomarkers may include ANGL6, HTRA1, PXDN, ANTR2, CSPG2, ANTR1, NOTUM, CILP1, CAN2, or GP1BB.The biomarkers may include ANGL6, HTRA1, PXDN, ANTR2, CSPG2, ANTR1, NOTUM, CILP1, CAN2, and GP1BB.

[0366] In some cases, the method includes assaying a plasma sample to detect the presence, absence, or abundance of one or more peptides or fragments of peptides from among the peptides listed in Table 2 or another table or figure provided herein. In some cases, the method includes assaying a buffy coat sample to detect the presence, absence, or abundance of one or more peptides or fragments of peptides from among the peptides listed in Table 2 or another table or figure provided herein. In some cases, the method includes assaying a granulocyte sample to detect the presence, absence, or abundance of one or more peptides or fragments of peptides from among the peptides listed in Table 2 or another table or figure provided herein. In some cases, the method includes assaying a homogenized tissue (e.g., a homogenized lung biopsy tissue sample) to detect the presence, absence, or abundance of one or more peptides or fragments of peptides from among the peptides listed in Table 2 or another table or figure provided herein.

[0367] The methods of the present disclosure enable rapid and deep biomolecular profiling from complex biological samples. In many cases, the methods detect and identify hundreds or thousands of distinct biomolecules. Such extensive analysis allows for more in-depth profiling of complex samples and increases the diagnostic utility of individual peptides. The methods of the present disclosure may include assaying a sample from a subject to detect the presence, absence, or abundance of at least 50 peptides from the biological sample with one or more additional peptides or peptide fragments from among the peptides listed in Table 2 or another table or figure provided herein. The methods of the present disclosure may include assaying a sample from a subject to detect the presence, absence, or abundance of at least 100 peptides from the biological sample with one or more additional peptides or peptide fragments from among the peptides listed in Table 2 or another table or figure provided herein. The methods of the present disclosure may include assaying a sample from a subject to detect the presence, absence, or abundance of at least 200 peptides from the biological sample with one or more additional peptides or peptide fragments from among the peptides listed in Table 2 or another table or figure provided herein. The disclosed methods may include assaying a sample from a subject to detect the presence, absence, or abundance of at least 400 peptides from the biological sample with one or more additional peptides or fragments of peptides from among the peptides listed in Table 2 or another table or figure provided herein. The disclosed methods may include assaying a sample from a subject to detect the presence, absence, or abundance of at least 600 peptides from the biological sample with one or more additional peptides or fragments of peptides from among the peptides listed in Table 2 or another table or figure provided herein. The disclosed methods may include assaying a sample from a subject to detect the presence, absence, or abundance of at least 800 peptides from the biological sample with one or more additional peptides or fragments of peptides from among the peptides listed in Table 2 or another table or figure provided herein.The disclosed method may include assaying a sample from a subject to detect the presence, absence, or abundance of at least 1,000 peptides from the biological sample with one or more additional peptides or peptide fragments from among the peptides listed in Table 2 or another table or figure provided herein. The disclosed method may include assaying a sample from a subject to detect the presence, absence, or abundance of at least 1,200 peptides from the biological sample with one or more additional peptides or peptide fragments from among the peptides listed in Table 2 or another table or figure provided herein. The disclosed method may include assaying a sample from a subject to detect the presence, absence, or abundance of at least 1,400 peptides from the biological sample with one or more additional peptides or peptide fragments from among the peptides listed in Table 2 or another table or figure provided herein. The disclosed method may include assaying a sample from a subject to detect the presence, absence, or abundance of at least 1,600 peptides from the biological sample with one or more additional peptides or peptide fragments from among the peptides listed in Table 2 or another table or figure provided herein. The disclosed methods may include assaying a sample from a subject to detect the presence, absence, or abundance of at least 1800 peptides from the biological sample along with one or more additional peptides or fragments of peptides from among the peptides listed in Table 2 or another table or figure provided herein. The disclosed methods may include identifying abundance or signal intensity (e.g., mass spectrometry signal intensity) ratios between at least a subset of at least 50, at least 100, at least 200, at least 400, at least 600, at least 800, at least 1000, at least 1200, at least 1400, at least 1600, or at least 1800 peptides from among the peptides listed in Table 2 or another table or figure provided herein and the one or more additional peptides or fragments of peptides.

[0368] The disclosed methods may include monitoring the progression of lung cancer over time. The disclosed methods may include monitoring lung nodules over time. The methods may include collecting two samples from a patient at two different time points and detecting at least two peptides in each of the samples from among the peptides listed in Table 2 or another table or figure provided herein. The methods may include collecting two samples from a patient at two different time points and detecting at least three peptides in each of the samples from among the peptides listed in Table 2 or another table or figure provided herein. The methods may include collecting two samples from a patient at two different time points and detecting at least four peptides in each of the samples from among the peptides listed in Table 2 or another table or figure provided herein. The methods may include collecting two samples from a patient at two different time points and detecting at least five peptides in each of the samples from among the peptides listed in Table 2 or another table or figure provided herein. The method may include collecting two samples from a patient at two different time points and detecting in each of the samples at least six peptides from among the peptides listed in Table 2 or another table or figure provided herein. The method may include collecting two samples from a patient at two different time points and detecting in each of the samples at least seven peptides from among the peptides listed in Table 2 or another table or figure provided herein. The method may include collecting two samples from a patient at two different time points and detecting in each of the samples at least eight peptides from among the peptides listed in Table 2 or another table or figure provided herein. The method may include collecting two samples from a patient at two different time points and detecting in each of the samples at least nine peptides from among the pepti...

Claims

1. A multi-omics method comprising: sequencing nucleic acid obtained from the biological fluid sample obtained from the subject to generate nucleic acid sequence data; Quantifying the nucleic acid to obtain a nucleic acid measurement; processing proteins from the biological fluid sample to obtain proteomic data; generating a combined data set comprising at least a portion of the proteomic data, at least a portion of the nucleic acid measurements, and at least a portion of the nucleic acid sequence data; determining the presence or absence of cancer, at least in part, by inputting the combined dataset into a trained machine learning classifier, the trained machine learning classifier having a performance characteristic including an area under the curve (AUC) of a receiver operating characteristic (ROC) curve of at least 0.83, the trained machine learning classifier being trained on a first sample of cancer samples and a non-cancer sample. A method comprising:

2. The method of claim 1, further comprising contacting the biological fluid sample with nanoparticles, wherein the protein adsorbs to the nanoparticles to form a biomolecular corona.

3. The method of claim 1 or 2, wherein processing the protein includes performing mass spectrometry or immunoassay.

4. The method of claim 2, wherein the nanoparticles comprise a group of physiochemically distinct nanoparticles.

5. The method of claim 4, wherein the group of physiochemically distinct nanoparticles comprises lipid nanoparticles, metal nanoparticles, silica nanoparticles, or polymer nanoparticles.

6. The method of claim 1, wherein the nucleic acid comprises mRNA and / or miRNA.

7. The method of claim 1, wherein the nucleic acid comprises DNA.

8. The method of claim 1, wherein the trained machine learning classifier distinguishes between the presence and absence of cancer with at least 4% greater sensitivity at a predetermined specificity than a trained machine learning classifier trained using only proteomic data or trained using only nucleic acid sequence information and nucleic acid abundance, wherein the sensitivity of the trained machine learning classifier is determined in a dataset derived from a study of more than 20 subjects with cancer and more than 20 control subjects without the cancer.

9. The method of claim 1, wherein the AUC is determined by application of the trained machine learning classifier when applied to a dataset derived from a study of at least 20 subjects with cancer and more than 20 control subjects without cancer.

10. The determining step comprises: generating a first label corresponding to the presence, absence, or likelihood of the cancer based on the proteomic data; generating a second label corresponding to the presence, absence, or likelihood of the cancer based on the nucleic acid sequence information and the amount of the sequenced nucleic acid; (a), (b), or (c): (a) generating an unweighted average of the first and second labels, the first and second labels comprising a probability of the presence or absence of cancer; (b) generating a weighted average of the first and second labels, the first and second labels comprising a probability of the presence or absence of cancer; or (c) generating a majority score based on the first and second labels, the first and second labels comprising an indication of the likelihood of cancer. To carry out The method of claim 1 , comprising:

11. The method of claim 10, further comprising assigning weights to the first and second labels based on area under a ROC curve, area under a precision-recall curve, precision, precision, recall, sensitivity, F1-score, specificity, or a combination thereof.

12. The method of claim 11, identifying a subset of top features from the proteomic data based on feature importance in a test dataset; identifying at least a subset of top features from the nucleic acid sequence information and the nucleic acid quantification information based on feature importance in a test dataset, where the feature importance is determined at least in part based on a machine learning model; pooling a subset of the top features from the proteomic data and at least a subset of the top features from the nucleic acid sequence information and the amount of sequenced nucleic acid; The method of claim 1 further comprising:

13. The method of claim 12, wherein identifying the subset of top features from the proteomic data comprises identifying protein features from univariate protein data that have improved classification performance compared to other protein features in the univariate protein data, or wherein identifying the subset of at least the top features from the nucleic acid sequence information and the amount of sequenced nucleic acid comprises identifying nucleic acid features from univariate nucleic acid data that have improved classification performance compared to other nucleic acid features in the univariate nucleic acid data.

14. The method of claim 12, wherein the feature importance is determined by evaluating a change in performance of the machine learning model when a feature in the set of features is not present.

15. The method of claim 1, wherein the trained machine learning classifier is trained using deep learning, hierarchical cluster analysis, principal component analysis, partial least squares discriminant analysis, random forest classification analysis, support vector machine analysis, k-nearest neighbor analysis, naive Bayes analysis, K-means clustering analysis, gradient boosting analysis, boosted decision tree analysis, or hidden Markov analysis.

16. The method described in claim 1, wherein the cancer is selected from the group consisting of lung cancer, pancreatic cancer, colon cancer, liver cancer, breast cancer and ovarian cancer.

17. The method of claim 1, wherein the biological fluid sample comprises a blood, serum, or plasma sample.

18. The method of claim 1, wherein the subject is a human.

19. The method of claim 1, wherein the trained machine learning classifier comprises performance characteristics including an average AUC of the ROC curve of at least 0.

85.

20. The method of claim 1, further comprising classifying the biological fluid sample as having the presence or absence of cancer.

Images