Estimation of tumor purity from a single sample

Abstract

To provide a method for estimating tumor purity from a tumor sample without usage of matched-normal control.SOLUTION: A method is configured to: identify a set of genomic regions on the basis of nucleic acid sequence data that is aligned to a reference genome, in which each genomic region of the set of genomic regions includes one or more nucleotide-sequence variants related to a corresponding genomic region of the reference genome,; measure a B-allele frequency distribution of a biological sample on the basis of a B-allele frequency measured with respect to each genomic region of the set of genomic regions; process the B-allele frequency distribution using a trained machine-learning model; and estimate a metric identifying tumor purity in the biological sample.SELECTED DRAWING: None

Description

[Technical field]

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS This application claims priority to U.S. Provisional Patent Application No. 62 / 931,096, filed November 5, 2019, which is incorporated by reference in its entirety for all purposes.

[0002] The present disclosure relates generally to systems and methods for estimating tumor purity from a single sample, and more particularly, but not by way of limitation, the present disclosure relates to estimating tumor purity of a biological sample by processing B allele frequency distributions using a trained machine learning model. [Background technology]

[0003] Tumor cellularity, also referred to as "tumor purity", determines the proportion of cancer cells in a sample. Accurate estimation of tumor purity in a biological sample can contribute to improved accuracy in detecting changes in the amount and / or copy number of somatic mutations. This is because tumor purity indicates the allele frequency of somatic mutations present in a biological sample. Detection of somatic mutations and copy number variations can then be used to determine the stage of a subject's cancer or evaluate whether a particular cancer treatment is effective. Thus, tumor purity can be useful in determining the stage of cancer and / or evaluating the efficacy of treatment.

[0004] Although tumor purity can be an effective metric, it can also be a confounding variable in some bioinformatics analyses. For example, traditional approaches to estimating tumor purity may require histopathological evaluation by a pathologist by manually inspecting images of the sample to estimate tumor purity. However, histopathological evaluation, including manual inspection of sample images, can be subjective and inaccurate. Other traditional techniques for estimating tumor purity require comparing values ​​(e.g., putative somatic mutations) derived from nucleic acid sequence data of a given tumor sample with other values ​​derived from nucleic acid sequence data of a matched normal control sample. However, such normal control samples may not be available.

[0005] For example, conventional techniques estimate the tumor purity of a sample as a function of the allele fraction of somatic mutations unique to an individual's tumor. In the absence of a matched normal sample, identification of these somatic mutations is less accurate, greatly reducing the accuracy of the estimated purity. In some cases, a matched normal control is not available if (for example) the sample source did not collect or sequence a normal control.

[0006] Therefore, there is a need to accurately estimate tumor purity in a sample to facilitate detection without relying on subjective analysis (eg, histopathological evaluation) or the presence of normal control samples. Summary of the Invention [Problem to be solved by the invention]

[0007] In some embodiments, a method of estimating tumor purity is provided. The method may include obtaining nucleic acid sequence data representing a plurality of nucleic acid molecules of a tumor sample of a subject. The method may also include aligning the nucleic acid sequence data to a reference genome. The method may also include identifying a set of genomic regions based on the aligned nucleic acid sequence data. In some cases, each genomic region of the set of genomic regions comprises one or more nucleotide sequence variants associated with a corresponding genomic region of the reference genome.

[0008] The method may also include measuring a B allele frequency for each genomic region of the set of genomic regions. The method may also include measuring a B allele frequency distribution for the biological sample based on the B allele frequencies for the set of genomic regions. The method may also include processing the B allele frequency distribution using a trained machine learning model to estimate a metric identifying tumor purity for the biological sample. The method may also include outputting the metric.

[0009] In some embodiments, a system is provided that includes one or more data processors and a non-transitory computer-readable storage medium that includes instructions that, when executed on the one or more data processors, cause the one or more data processors to perform some or all of one or more methods disclosed herein.

[0010] In some embodiments, a computer program product is provided that is tangibly embodied in a non-transitory machine-readable storage medium and includes instructions configured to cause one or more data processors to perform some or all of one or more of the methods disclosed herein.

[0011] Some embodiments of the present disclosure include a system including one or more data processors. In some embodiments, the system includes a non-transitory computer-readable storage medium including instructions that, when executed on the one or more data processors, cause the one or more data processors to perform some or all of one or more methods and / or some or all of one or more processes disclosed herein. Some embodiments of the present disclosure include a computer program product tangibly embodied in a non-transitory machine-readable storage medium and including instructions configured to cause one or more data processors to perform some or all of one or more methods and / or some or all of one or more processes disclosed herein.

[0012] The terms and expressions used are used as terms of description and not of limitation, and there is no intention to use such terms and expressions to exclude any equivalents of the features shown and described or portions thereof, but it is recognized that various modifications are possible within the scope of the invention as claimed. Thus, while the invention as claimed has been specifically disclosed by embodiments and desired features, it is to be understood that modifications and variations of the concepts disclosed herein may be made by those skilled in the art. Such modifications and variations are deemed to be within the scope of the invention as defined by the appended claims. [Brief description of the drawings]

[0013] The features, embodiments, and advantages of the present disclosure will be better understood upon reading the following detailed description in conjunction with the following figures.

[0014] [Figure 1] FIG. 1 shows a schematic diagram of a fully connected neural network for estimating tumor purity according to some embodiments. [Diagram 2] FIG. 2 shows a schematic diagram of a one-dimensional convolutional neural network for estimating tumor purity according to some embodiments. [Diagram 3] FIG. 3 shows a schematic diagram of a two-dimensional convolutional neural network for estimating tumor purity according to some embodiments. [Figure 4A] FIG. 4A shows loss versus epoch plots for a fully connected neural network, a one-dimensional convolutional neural network, and a two-dimensional convolutional neural network trained on a VCF-level dataset, according to some embodiments. [Figure 4B] FIG. 4B shows root mean square error (RMSE) and mean absolute error (MAE) data for a fully connected neural network, a one-dimensional convolutional neural network, and a two-dimensional convolutional neural network trained on the VCF-level dataset, according to some embodiments. [Figure 4C]FIG. 4C shows root mean square error (RMSE) and mean absolute error (MAE) data for testing of the FASTQ level dataset, according to some embodiments. [Figure 5A] FIG. 5A shows a plot 500A comparing true tumor purity with tumor purity predicted by a trained fully connected neural network, according to some embodiments. [Figure 5B] FIG. 5B shows a plot 500B comparing true tumor purity to tumor purity predicted by a trained two-dimensional convolutional neural network, according to some embodiments. [Figure 5C] FIG. 5C shows a plot 500C comparing true tumor purity to tumor purity predicted by a one-dimensional convolutional neural network, according to some embodiments. [Figure 6] FIG. 6 shows an example schematic for estimating tumor purity using a trained machine learning model, according to some embodiments. [Figure 7A] 7A-7E provide examples of plotted BAF distributions according to some embodiments. [Figure 7B] 7A-7E provide examples of plotted BAF distributions according to some embodiments. [Figure 7C] 7A-7E provide examples of plotted BAF distributions according to some embodiments. [Figure 7D] 7A-7E provide examples of plotted BAF distributions according to some embodiments. [Figure 7E] 7A-7E provide examples of plotted BAF distributions according to some embodiments. [Figure 8] FIG. 8 shows a plot 800 of the first two principal components for classifying biological samples using B allele frequency features, according to some embodiments. [Figure 9] FIG. 9 includes a flow chart illustrating an example of a method for estimating tumor purity of a biological sample according to certain embodiments. [Figure 10]FIG. 10 illustrates an example computer system for implementing some of the embodiments disclosed herein. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0015] I. Overview To address at least the above deficiencies of conventional systems, the approach of the present invention can be used to estimate tumor purity by processing the B allele frequency distribution of heterozygous sites of sequencing data using a trained machine learning model. The trained machine learning model can generate an estimated metric that identifies the tumor purity of a biological sample even in the absence of matching sequencing data of a normal control sample. Thus, the technology of the present invention can accurately estimate the tumor purity in a biological sample without a matched normal control, thereby facilitating more accurate analysis of nucleic acid sequencing data from tumor-only samples.

[0016] Nucleic acid sequence data representing a plurality of nucleic acid molecules of a tumor sample of a subject can be obtained. In some embodiments, the nucleic acid sequence data is whole-exome sequencing data. The nucleic acid sequence data may be whole-genome sequencing data. In some embodiments, the sequencing data is from a tumor sample. The tumor sample may be from a human subject. The nucleic acid sequence data may be generated by shotgun sequencing. In some embodiments, the nucleic acid sequence data may be generated by sequencing a selected portion of the genome or exome.

[0017] The nucleic acid sequence data may be aligned to a reference genome. A set of genomic regions may be identified based on the aligned nucleic acid sequence data. In some cases, each genomic region of the set of genomic regions comprises one or more nucleotide sequence variants associated with the corresponding genomic region of the reference genome. To identify the nucleotide sequence variants, variant candidates may be identified and the reference read depth and alternative read depth of the variant candidates may be calculated.

[0018] The B-allele frequency of each genomic region of the set of genomic regions can be measured. B-allele frequency (BAF) refers to a normalized measure of the allele intensity ratio of two alleles (A and B), where a BAF of 1 or 0 indicates the complete absence of one of the two alleles (e.g., AA or BB), and a BAF of 0.5 indicates that both alleles (e.g., AB) are present equally. Based on the B-allele frequency of the set of genomic regions, the B-allele frequency distribution of the biological sample can be measured. In some embodiments, the B-allele frequency distribution is normalized.

[0019] The trained machine learning model may be used to process the B allele frequency distribution to estimate a metric that identifies the tumor purity of the biological sample. As used herein, tumor purity or tumor cell content refers to the percentage of cancer cells in a tumor sample. Tumor purity may be a metric used in various approaches to estimate tumor features that correspond to direct clinical relevance. For example, the estimated tumor purity may directly change the threshold of evidence required to identify copy number alterations, including those associated with approved companion diagnostics (e.g., EGFR exon 19 deletion in NSCLC). Tumor purity estimates may also be used to improve the quality (sensitivity, PPV) of somatic variant calls by prioritizing the expected allele fraction of somatic variants. For example, a tumor purity value of 60% may indicate that a tumor sample contains 60% tumor cells and 40% normal cells. Additionally or alternatively, tumor purity may indicate the number of cancer cells in a tumor sample. Note that whenever the term "tumor purity" is mentioned, the term "tumor cell content" may be used interchangeably throughout this specification.

[0020] The trained machine learning model may be trained on a dataset generated in silico (e.g., nucleic acid sequence data generated by computer simulation). In some embodiments, the trained machine learning model is trained on a training dataset generated from sequencing data derived from tumor cells diluted with normal cells. The mean absolute error of the trained machine learning model may be less than about 0.2. In some embodiments, the root mean square error of the trained machine learning model is less than about 0.2.

[0021] In some embodiments, the trained machine learning model includes a fully connected neural network. The fully connected network may include a fully connected layer with a Rectified Linear Unit (ReLU) activation function. In some embodiments, an output activation function of the fully connected neural network is a sigmoid function. A loss function of the fully connected neural network may be configured to calculate a mean squared error (MSE). In some embodiments, the fully connected neural network is tuned by a hyperparameter search using random sampling with a linear search of layers, a linear search of size, a logarithmic search of learning rate, or a combination thereof.

[0022] In some cases, the trained machine learning model includes a one-dimensional convolutional neural network that may be configured to use the B allele frequency distribution as an input and encode an input size of height 25, width 1, and depth 100. In some embodiments, each layer of the one-dimensional convolutional neural network performs a 1×1 convolution followed by a ReLU activation function.

[0023] In some embodiments, the trained machine learning model includes a two-dimensional convolutional neural network. In some embodiments, the B allele frequency distribution of the two-dimensional convolutional neural network is used as an input and encoded into an input size defined as height 25, width 100, and depth 1. In some embodiments, each convolutional layer of the two-dimensional convolutional neural network is followed by a ReLU layer. The output of the two-dimensional convolutional neural network may be a densely connected layer with a sigmoid activation function. Additionally, the hyperparameters of the two-dimensional convolutional neural network may be tuned by adjusting the number of layers, the filter size, the number of filters, or a combination thereof.

[0024] An estimated metric identifying the tumor purity may be output. For example, a report including the estimated metric may be output. In some embodiments, the report includes information identifying a B allele frequency distribution. The report may also include information identifying at least one diagnostic marker and / or at least one prognostic marker. In some embodiments, the report includes information identifying a predicted somatic variant. The report may also include a treatment recommendation. For example, the estimated tumor purity may indicate an elevated methylation level of O-6-methylguanine-DNA methyltransferase, a prognostic biomarker for glioblastoma. In another example, the estimated tumor purity may be used to measure the amount of mutational burden of the cancer cells, which may be used to determine a particular type of immunotherapy. Depending on the amount of estimated tumor purity, a recommendation may be made as to whether or not to initiate a treatment for glioblastoma. In some embodiments, the treatment recommendation may include a recommendation to administer a treatment to the human subject. The treatment recommendation may include a recommendation not to administer a treatment to the human subject.

[0025] Thus, embodiments of the present disclosure provide technical advantages over conventional systems by providing a more accurate estimate of tumor purity that does not have to rely on data from matched normal control samples. The estimated tumor purity may be used to improve the accuracy of analysis and annotation of sequencing data from tumor samples. One or more reports describing the estimated tumor purity may be generated (e.g., diagnostic and / or prognostic reports). For example, estimates of tumor purity may be used to improve the accuracy of diagnostic techniques that identify somatic mutations and / or copy number alterations, and a report may be generated that includes details of the predicted somatic mutations and / or copy number alterations.

[0026] Based on the tumor purity estimate and / or the report facilitated by the tumor purity estimate, one or more treatments can be administered to the patient or withheld from the patient. For example, the predicted somatic variants can be compared to one or more databases of known cancer mutations to diagnose or characterize the cancer. Variants associated with responsiveness or non-responsiveness to a particular cancer treatment can be identified, and treatment recommendations can be provided. The cancer can be treated based on the recommendations.

[0027] The following examples are provided to introduce certain embodiments. In the following description, for the purpose of explanation, specific details are set forth to provide a thorough understanding of the examples in the present disclosure. However, it is clear that various examples can be practiced without these specific details. For example, devices, systems, structures, assemblies, methods, and other components may be shown as components in block diagram form so as not to obscure the examples in unnecessary detail. In other examples, well-known devices, processes, systems, structures, and techniques may be shown without necessary details so as to avoid obscuring the examples. The drawings and descriptions are not limiting. The terms and expressions used in this disclosure are used as terms of description and not of limitation, and there is no intention to use such terms or expressions to exclude any equivalents of the features shown and described or portions thereof. The word "example" is used herein to mean "serve as an example, illustration, or illustration." Any embodiment or design described herein as an "example" is not necessarily to be construed as preferred or advantageous over other embodiments or designs.

[0028] II. Machine learning model for estimating tumor purity from a single sample A. Example of a machine learning model for estimating tumor purity of a biological sample As described herein, the tumor purity of a biological sample can be estimated using a trained machine learning model. The trained machine learning model may correspond to one of various machine learning models trained to estimate the tumor purity of a biological sample. In some embodiments, the trained machine learning model includes multiple models (e.g., 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 machine learning models). For example, one of three machine learning models, such as a fully connected neural network, a one-dimensional convolutional neural network, and a two-dimensional convolutional neural network, can be trained to estimate the tumor purity in nucleic acid sequencing data. In some cases, the trained machine learning model includes a deep neural network. Deep neural networks can be used to capture the internal structure of an expanding high-dimensional dataset (e.g., nucleic acid sequence data). Deep neural networks identify high-level features, improving performance over traditional statistical models, increasing interpretability, and providing further understanding of the structure of nucleic acid sequence data.

[0029] A trained machine learning model may include hyperparameters. Hyperparameters may be configured to be external to the model and have values ​​not estimated from data (e.g., training data or input data). In some cases, hyperparameters are tuned, e.g., tuned to solve a particular predictive modeling problem. In some cases, hyperparameters are used to help estimate model parameters. Hyperparameters may be user specified. In some cases, hyperparameters may be determined using a set of heuristic algorithms.

[0030] FIG. 1 shows a schematic diagram 100 of a fully connected neural network for estimating tumor purity according to some embodiments. The input layer 105 of the fully connected neural network is shown on the left, followed by a set of hidden layers 110. The output layer 115 is shown on the right. For the fully connected network, the input features can include the whole exome, B allele frequency (BAF) distribution of the nucleotide sequence variants identified in the biological sample. The fully connected neural network may include a series of fully connected layers with a ReLU activation function, and the output activation function can be a sigmoid function. In some cases, the loss function of the fully connected neural network is configured to generate a mean squared error (MSE). The fully connected neural network can be optimized by hyperparameter search using random sampling, such as a linear search of layers and sizes and a logarithmic search of the learning rate. Each output dimension of the fully connected neural network can depend on each input dimension. In some cases, the fully connected neural network can be a feed-forward neural network.

[0031] Convolutional neural networks can be trained to estimate the tumor purity of biological samples. Convolutional neural networks rely on local connections and combined weights between units, followed by feature pooling (subsampling) to obtain translation-invariant descriptors. A basic convolutional neural network architecture includes one convolutional layer and a pooling layer, optionally followed by a fully connected layer for supervised prediction. In some cases, convolutional neural networks are composed of multiple (e.g., >10) convolutional and pooling layers to better model the input space. Convolutional neural networks may require large datasets to be adequately trained. In some embodiments, convolutional neural networks use fewer parameters than fully connected neural networks by computing convolutions in small regions of the input space and sharing parameters between regions. The convolutional neural network can be a one-dimensional convolutional neural network. The convolutional neural network can be a two-dimensional convolutional neural network. In some embodiments, the convolutional neural network has three or more dimensions.

[0032] FIG. 2 shows a schematic diagram 200 of a one-dimensional convolutional neural network for estimating tumor purity according to some embodiments. For the one-dimensional convolutional neural network, the input function may include a chromosomal BAF distribution and may be encoded into an input size that defines a height of 25, a width of 1, and a depth of 100. Each layer of the one-dimensional convolutional neural network may perform a 1×1 convolution followed by a ReLU activation function. In some cases, a "network within a network" deep network structure is used. In effect, one-dimensional (1D) convolutions with a stride of the depth of each layer on a flattened input are used. In some cases, hyperparameters corresponding to the depth and layer size are obtained from a fully connected neural network (see FIG. 1), and a hyperparameter corresponding to the learning rate of the one-dimensional convolutional neural network may be tuned for optimization.

[0033] FIG. 3 shows a schematic diagram 300 of a two-dimensional convolutional neural network for estimating tumor purity according to some embodiments. For the two-dimensional convolutional neural network, the input features may include chromosomal BAF distributions and may be encoded into an input size that defines a height of 25, a width of 100, and a depth of 1. Each layer corresponds to a two-dimensional convolutional layer (a two-dimensional convolutional neural network layer) followed by a ReLU activation function. The output layer may correspond to a densely connected layer with a sigmoid activation function. In some embodiments, the hyperparameters of the two-dimensional convolutional neural network are optimized by searching for the number of layers, the filter size, and the number of filters. Dropout and / or L2 regularization can be performed to reduce initial overfitting of the training set.

[0034] Other types of machine learning models can be trained to estimate the tumor purity of the biological sample. In some embodiments, the machine learning model corresponds to one or more of gradient boosting decision trees (e.g., XGBoost framework or LightGBM framework), bagging procedures, boosting procedures, and / or random forest algorithms. For example, a gradient boosted decision tree can be trained to estimate the tumor purity of the biological sample. Gradient boosting corresponds to a type of machine learning technique that can be used for regression and classification problems, as well as for generating predictive models that may include an ensemble of weak predictive models, such as decision trees. In some cases, the gradient boosted decision tree can include, for example, the XGBoost framework or the LightGBM framework.

[0035] B. Training Dataset for Training Machine Learning Models to Estimate Tumor Purity of Biological Samples The machine learning model for estimating tumor purity of a biological sample can be trained using a supervised training algorithm. A supervised training algorithm may be used to train the machine learning model and generate an output corresponding to a linear regression value and / or a logistic regression value corresponding to tumor purity. The machine learning model can be trained based on a training dataset. In some cases, the training dataset includes sequencing datasets from pure tumor samples, pure normal samples, and a dilution series of tumor cells among normal cells. The sequencing datasets may be derived from biological samples of other subjects, including biological samples from subjects diagnosed with cancer. The biological samples of the training dataset may correspond to malignant tissue, benign tissue, or a mixture thereof. In some embodiments, the biological sample including the cancer tissue is obtained without a matching normal sample. Additionally or alternatively, a matching normal sample is obtained (for example) for training and testing the machine learning model.

[0036] In some embodiments, the training dataset comprises a training nucleotide sequence variant dataset. The training variant dataset may correspond to nucleic acid sequence data in which nucleotide sequence variants are identified. The variants following the nucleotide may include single nucleotide polymorphisms (SNPs), one or more single nucleotide changes, insertion-deletion mutations (indels), small insertions, small deletions, structural variant junctions, variable length tandem repeats, flanking sequences, and combinations thereof. In some cases, biased "Variant Call Format (VCF) level" datasets are generated using an artificial combination of VCF files generated from training samples including tumor and normal samples. Thus, the training dataset may include in silico "dilutions" of tumor sequences. VCF level datasets may be generated with the characteristics outlined in Table 1. To train a machine learning model, the machine learning model may be fitted to a "train" set, hyperparameters may be tuned based on performance on the "train" set and the "dev" set, and final performance may be evaluated on a "test" set. [Table 1]

[0037] In some embodiments, the training dataset includes training nucleic acid sequence data. The training nucleic acid sequence data may correspond to an unbiased "FASTQ-level" dataset generated by downsampling and combining reads from training samples. The training samples may include normal and tumor samples. The FASTQ-level dataset can be generated with sizes 99 and 11 unique samples and may be obtained from sequencing in vitro dilutions of tumor cells and matched normal cells. Compared to the VCF-level dataset, the FASTQ-level dataset is closer to the raw data because it may be at the read level rather than the variant level. In some cases, the FASTQ-level dataset is used to improve or optimize performance using more realistic data. Since it is computationally difficult to generate large FASTQ-level datasets, machine learning models may be trained using the FASTQ-level dataset using transfer learning. Transfer learning is performed using machine learning models pre-trained on the VCF-level dataset, with 5-fold cross-validation. In some cases, each machine learning model may be trained on 4 / 5 of the training dataset and tested on 1 / 5 of the training dataset at each iteration.

[0038] C. Training and testing a machine learning model to estimate tumor purity of a biological sample A machine learning model (e.g., a fully connected neural network) can be trained on a training dataset (e.g., a VCF-level dataset). In some cases, the machine learning model is first trained on a VCF-level dataset. The machine learning model can be fit to a "train" set, hyperparameters can be tuned based on performance on the "train" and "dev" sets, and final performance can be evaluated on a "test" set. FIG. 4A illustrates loss versus epoch plots 400A for a fully connected neural network, a one-dimensional convolutional neural network, and a two-dimensional convolutional neural network trained on a VCF-level dataset, according to some embodiments. The loss versus epoch plots 400A show that the performance level of each trained machine learning model increases with each epoch, thereby reaching an error value (e.g., mean squared error) in the range of approximately 0.01 to 0.025.

[0039] FIG. 4B shows root mean square error (RMSE) and mean absolute error (MAE) data 400B of fully connected neural networks, one-dimensional convolutional neural networks, and two-dimensional convolutional neural networks trained on the VCF level dataset, according to some embodiments. All models perform comparably on the VCF level dataset. For this dataset, the root mean square error values ​​are in the range of 0.081 (two-dimensional convolutional neural networks) to 0.088 (fully connected neural networks) based on the training dataset. As shown in FIG. 4A and FIG. 4B, the consistent performance of all machine learning models trained on the VCF level training and testing datasets indicates good fit and may be approaching optimality for estimating the tumor purity of a sample.

[0040] In some embodiments, the FASTQ level dataset is used to improve or optimize the performance of a trained machine learning model. Since it is computationally challenging to generate a large FASTQ level dataset, a machine learning model may be trained using the FASTQ level dataset using transfer learning. Transfer learning may be performed using a machine learning model pre-trained on the VCF level dataset. The performance of the trained machine learning model may be evaluated using a 5-fold cross-validation strategy. In some cases, each machine learning model may be trained on 4 / 5 of the training dataset and tested on 1 / 5 of the training dataset at each iteration. FIG. 4C shows RSME and MAE data 400C of a fully connected neural network, a one-dimensional convolutional neural network, and a two-dimensional convolutional neural network trained on the FASTQ level dataset, according to some embodiments. The RMSE and MAE data 400C show comparable levels of performance between the trained one-dimensional convolutional neural network (e.g., 0.057 RMSE) and the trained two-dimensional convolutional neural network (e.g., 0.067 RMSE). Furthermore, the RMSE and MAE data 400C show that the trained machine learning model achieved low RSME and MAE after transfer learning on the FASTQ-level dataset.

[0041] The performance level of each trained machine learning model was evaluated. Figure 5A shows a plot 500A comparing the true tumor purity to the tumor purity predicted by a trained fully connected network, according to some embodiments. Figure 5B shows a plot 500B comparing the true tumor purity to the tumor purity predicted by a trained two-dimensional convolutional neural network, according to some embodiments. Figure 5C shows a plot 500C comparing the true tumor purity to the tumor purity predicted by a one-dimensional convolutional neural network, according to some embodiments. The results from Figures 5A, 5B, and 5C, respectively, show that the trained machine learning models achieve a performance level comparable to existing tools that require matched normal control samples.

[0042] In some cases, convolutional neural network models tend to overfit to training data sets. For example, the higher RSME and MAE values ​​corresponding to convolutional neural networks may not be significantly improved by regularization and dropout. The tendency of these convolutional neural networks to overfit may indicate the ability of these machine learning models to detect complex signals of tumor purity when trained using larger training data sets. Therefore, the machine learning models may be further trained using larger and more diverse training data sets.

[0043] In some embodiments, the trained machine learning model is trained using the training data and / or the test data to achieve one or more pre-defined performance levels for estimating tumor purity. The one or more pre-defined performance levels include: Up to 1%, 2%, 3%, 4%, 5%, 6%, 7%, 8%, 9%, 10%, 11%, 12%, 13%, 14%, 15%, 16%, 17%, 18%, 19%, or 20% MAE. Maximum RMSE of 1%, 2%, 3%, 4%, 5%, 6%, 7%, 8%, 9%, 10%, 11%, 12%, 13%, 14%, 15%, 16%, 17%, 18%, 19%, or 20%. Accuracy of at least about 0.2, 0.25, 0.3, 0.35, 0.4, 0.45, 0.5, 0.55, 0.6, 0.65, 0.7, 0.75, 0.8, 0.85, 0.9, 0.95, or more. Additionally or alternatively, the trained machine learning model has an accuracy of about 0.2-1.0, 0.2-0.9, 0.2-0.8, 0.2-0.7, 0.2-0.6, 0.2-0.5, 0.2-0.4, 0.2-0.3, 0.3-1.0, 0.3-0.9, 0.3-0.8, 0.3-0.7, 0.3-0.6, 0.3-0.5, 0.3-0.4, 0.4-1.0, 0.4-0.9, 0.4-0. 8, trained to estimate tumor purity with an accuracy of 0.4-0.7, 0.4-0.6, 0.4-0.5, 0.5-1.0, 0.5-0.9, 0.5-0.8, 0.5-0.7, 0.5-0.6, 0.6-1.0, 0.6-0.9, 0.6-0.8, 0.6-0.7, 0.7-1.0, 0.7-0.9, 0.7-0.8, 0.8-1.0, 0.8-0.9, or 0.9-1.0. Additionally or alternatively, the trained machine learning model has a recall of at least about 0.2, 0.25, 0.3, 0.35, 0.4, 0.45, 0.5, 0.55, 0.6, 0.65, 0.7, 0.75, 0.8, 0.85, 0.9, 0.95, or more. Additionally or alternatively, the trained machine learning model has a recall of at least about 0.2-1.0, 0.2-0.9, 0.2-0.8, 0.2-0.7, 0.2-0.6, 0.2-0.5, 0.2-0.4, 0.2-0.3, 0.3-1.0, 0.3-0.9, 0.3-0.8, 0.3-0.7, 0.3-0.6, 0.3-0.5, 0.3-0.4, 0.4-1.0, 0.4-0.9, 0.4-0.8 , trained to estimate tumor purity with a recall of 0.4–0.7, 0.4–0.6, 0.4–0.5, 0.5–1.0, 0.5–0.9, 0.5–0.8, 0.5–0.7, 0.5–0.6, 0.6–1.0, 0.6–0.9, 0.6–0.8, 0.6–0.7, 0.7–1.0, 0.7–0.9, 0.7–0.8, 0.8–1.0, 0.8–0.9, or 0.9–1.0. An F1 score (e.g., macro-average F1 classification score) of at least about 0.2, 0.25, 0.3, 0.35, 0.4, 0.45, 0.5, 0.55, 0.6, 0.65, 0.7, 0.75, 0.8, 0.85, 0.86, 0.87, 0.88, 0.89, 0.9, 0.91, 0.92, 0.93, 0.94, 0.95, 0.96, 0.97, 0.98, 0.99, 0.995, or more. Additionally or alternatively, the trained machine learning model may have a frequency response of about 0.2-1.0, 0.2-0.99, 0.2-0.95, 0.2-0.9, 0.2-0.8, 0.2-0.7, 0.2-0.6, 0.2-0.5, 0.2-0.4, 0.2-0.3, 0.3-1.0, 0.3-0.99, 0.2-0.95, 0.3-0. 9, 0.3~0.8, 0.3~0.7, 0.3~0.6, 0.3~0.5, 0.3~0.4, 0.4~1.0, 0.4~0.99, 0.4~0.95, 0.4~0.9, 0.4~0.8, 0.4~0.7, 0.4~0.6, 0.4~0.5, 0.5~1.0, 0.5~0.99, 0.5~0.95, 0.5~0.9, 0.5~0.8, 0.5~0.7, 0.5~0.6, 0.6~1.0, 0.6~0.99, 0.6~0.95, 0.6~0.9, 0.6~0.8, 0.6~0.7, 0.7~1.0, 0.7~0.99, 0.7~0.98, 0.7~0.97, 0.7~0.96, 0.7~0.95, 0.7~0.9, 0.7~0. 8, trained to estimate tumor purity with an F1 score of 0.8–1.0, 0.8–0.99, 0.8–0.98, 0.8–0.97, 0.8–0.96, 0.8–0.95, 0.8–0.9, 0.9–1.0, 0.9–0.99, 0.9–0.98, 0.9–0.97, 0.9–0.96, or 0.9–0.95.

[0044] III. Estimation of Tumor Purity 6 shows an example of a schematic diagram 600 for estimating tumor purity using a trained machine learning model according to some embodiments. Tumor purity from sequencing of mismatched biological samples can be estimated using a trained machine learning model (e.g., a regression model). In some cases, a deep neural network is used to estimate tumor purity from the allele fraction distribution (e.g., B allele frequency distribution) of nucleotide sequence variants in unpaired biological samples. A regression model can be used to predict tumor purity from the B allele frequency (BAF) of whole exome sequencing data.

[0045] In block 605, a mismatched tumor sample can be obtained from a cancer patient (i.e., no matched normal sample). In block 610, DNA is extracted from the tumor sample, processed, and whole-exome sequenced. In some cases, the sequencing reads undergo quality control processing (e.g., via FastQC) to provide a FASTQ file. The FASTQ file is aligned to a reference genome to generate a BAM file.

[0046] In block 615, the GATK HaplotypeCaller can be used to call variants from the BAM file and generate a VCF file containing variant information. Heterozygous sites within the sample exome can be identified and the VCF file can be filtered to obtain reference and alternate read depths for all heterozygous sites. This information is used to calculate the B allele frequency (BAF). A normalized BAF distribution is calculated for heterozygous sites across the exome.

[0047] In block 620, the trained machine learning model can be used to estimate tumor purity from the normalized BAF distribution. Thus, the trained machine learning model can generate accurate estimates of tumor purity even in the absence of a matched normal sample.

[0048] A. Subjects and Samples To estimate tumor purity, nucleic acid sequence data representing a plurality of nucleic acid molecules can be obtained from a biological sample of a subject. The subject can be a human. The subject can be male or female. The subject can be a fetus, an infant, a child, an adolescent, a teenager, or an adult. The subject can be a patient of any age. For example, the subject can be a patient less than about 10 years old. For example, the subject can be a patient at least about 0, 5, 10, 20, 30, 40, 50, 60, 70, 80, 90, or 100 years old. The subject can be a patient or other individual undergoing or being evaluated for a therapeutic regimen (e.g., cancer treatment). However, in some cases, the subject is not undergoing a therapeutic regimen.

[0049] In some cases, the subject may be a mammal or a non-mammal. In some cases, the subject is a mammal, such as a human, a non-human primate (e.g., ape, monkey, chimpanzee), cat, dog, rabbit, goat, horse, cow, pig, rodent, mouse, SCID mouse, rat, guinea pig, or sheep. In some embodiments, species variants or homologs of these genes are used in the non-human animal model. Species variants may be genes of different species that have the greatest sequence identity and similarity in functional properties with each other. Many of such species variants of human genes may be listed in the Swiss-Prot database.

[0050] Certain embodiments may include obtaining a sample from a subject, such as a human subject. In some cases, a clinical specimen from a patient is obtained. For example, blood may be drawn from a patient. Certain embodiments may include specifically detecting, profiling, or quantifying molecules (e.g., nucleic acids, DNA, RNA, etc.) within a biological sample.

[0051] The sample may be a tissue sample or a bodily fluid. Optionally, the sample is a tissue sample or an organ sample such as a biopsy. Optionally, the sample comprises cancer cells. Optionally, the sample comprises cancerous cells and normal cells. Optionally, the sample is a tumor biopsy. The bodily fluid may be sweat, saliva, tears, urine, blood, menses, semen, and / or cerebrospinal fluid. Optionally, the sample is a blood sample. The sample may comprise one or more peripheral blood lymphocytes. The sample may be a whole blood sample. The blood sample may be a peripheral blood sample. Optionally, the sample comprises peripheral blood mononuclear cells (PBMCs), and optionally, the sample comprises peripheral blood lymphocytes (PBLs). The sample may be a serum sample.

[0052] The sample may be obtained using any method capable of providing a sample suitable for the analytical methods described herein. The sample may be obtained by non-invasive methods such as throat swabs, oral swabs, bronchial lavage, urine collection, skin or cervical scraping, buccal swabs, saliva collection, stool collection, menstrual blood collection, or semen collection. The sample may be obtained by minimally invasive methods such as blood collection. The sample may be obtained by venipuncture. In other examples, the sample is obtained by invasive procedures including, but not limited to, biopsy, alveolar or pulmonary lavage, or needle aspiration. Methods of biopsy may include surgical biopsy, incisional biopsy, excision biopsy, punch biopsy, shaved biopsy, or skin biopsy. The sample may be a formalin-fixed section. Methods of needle aspiration may further include fine needle aspiration, core needle biopsy, vacuum assisted biopsy, or large core biopsy. In some cases, multiple samples may be obtained by the methods herein to ensure a sufficient amount of biological material. In some cases, the sample is not obtained by biopsy. Optionally, the sample is not a kidney biopsy.

[0053] The methods of the present disclosure can be used to estimate tumor purity in samples that contain at least about 1%, 5%, 10%, 15%, 20%, 25%, 30%, 35%, 40%, 45%, 50%, 55%, 60%, 65%, 70%, 75%, 80%, 85%, 90%, 95% or more tumor cells. In some embodiments, the methods of the present disclosure can be used to estimate tumor purity in samples that contain up to about 10%, 15%, 20%, 25%, 30%, 35%, 40%, 45%, 50%, 55%, 60%, 65%, 70%, 75%, 80%, 85%, 90%, 95% or more tumor cells.

[0054] B. Generation of Nucleic Acid Sequencing Data In some embodiments, the sample is processed to obtain nucleic acid sequence data. A "nucleic acid" or "nucleic acid molecule" may correspond to a polymeric form of nucleotides of any length, either ribonucleotides, deoxyribonucleotides, or peptide nucleic acids (PNAs), containing purine and pyrimidine bases, or other natural, chemically or biochemically modified, non-natural, or derivatized nucleotide bases. The backbone of a polynucleotide may contain sugar and phosphate groups as typically found in RNA or DNA, or modified or substituted sugar or phosphate groups. A polynucleotide may contain modified nucleotides, such as methylated nucleotides and nucleotide analogs. The sequence of nucleotides may be interrupted by non-nucleotide components. Thus, the terms nucleoside, nucleotide, deoxynucleoside, and deoxynucleotide generally include analogs such as those described herein. These analogs are molecules that have some structural features in common with naturally occurring nucleosides or nucleotides, and therefore, when incorporated into a nucleic acid sequence or oligonucleoside sequence, allow hybridization with naturally occurring nucleic acid sequences in solution. Typically, these analogs are derived from naturally occurring nucleosides and nucleotides by substituting and / or modifying the base, ribose, or phosphodiester moiety. The changes can be tailored to stabilize or destabilize hybridization or to increase the specificity of hybridization with complementary nucleic acid sequences, as needed. The nucleic acid molecule can be a DNA molecule. The nucleic acid molecule can be an RNA molecule.

[0055] Sample processing includes nucleic acid sample processing followed by nucleic acid sample sequencing. Part or all of the biological sample can be sequenced to provide nucleic acid sequence data, which may be stored or maintained in an electronic, magnetic, or optical repository. The sequence information may be analyzed by a computer processor, and the analyzed sequence information may be stored in an electronic storage location. The electronic repository may contain a pool or collection of sequence information generated from the nucleic acid sample and analyzed sequence information. In some embodiments, the biological sample is obtained from a subject having or suspected of having cancer.

[0056] In some embodiments, nucleic acid sequencing data is generated from pure tumor and pure normal samples. Matched paired cell lines are available from separate sources (e.g., American Type Culture Collection (ATCC)). Each matched pair may include tumor and normal cell lines of the same subject. The cell lines may be cultured and expanded in vitro to obtain a suitable number of cells for DNA extraction. DNA is extracted, processed, and subjected to whole-exome or whole-genome sequencing. Sequence reads may be subjected to quality control processing (e.g., via FastQC) to provide FASTQ files.

[0057] In some cases, the nucleic acid sequence data is generated using whole genome sequencing. In some cases, whole genome sequencing is used to identify variants in an individual. In some cases, sequencing can include deep sequencing over a portion of the genome. For example, a fraction of the genome can be at least about 50, 75, 100, 125, 150, 175, 200, 225, 250, 275, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750, 800, 850, 900, 950, 1,000, 1100, 1200, 1300, 1400, 1500, 1600, 1700, 1800, 2000, 2200, 2300, 2400, 2500, 2600, 2750, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750, 800, 850, 900, 950, 1,000, 1100, 1200, 1300, 1400, 1500, 1600, 1700, 2000, 2500, 2600, 2750, 2800, 2900, 3000, 3100, 3200, 3300, 3400, 3500, 3600, 3700, 3800, 3900, 4000, 4100, 4200, The genome may be sequenced over 800, 1900, 2,000, 3,000, 4,000, 5,000, 6,000, 7,000, 8,000, 9,000, 10,000, 15,000, 20,000, 30,000, 40,000, 50,000, 60,000, 70,000, 80,000, 90,000, 100,000, or more bases or base pairs. In some cases, the genome may be sequenced over 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, or more than 10 million bases or base pairs. In some cases, the genome may be sequenced over the entire exome (e.g., whole exome sequencing). In some cases, deep sequencing can include obtaining multiple reads over a portion of a genome. For example, obtaining multiple reads can include at least 2, 3, 4, 5, 6, 7, 8, 9, 10, 20, 30, 40, 50, 60, 70, 80, 90, 100, 200, 300, 400, 500, 600, 700, 800, 900, 1000, 10,000 or more reads over a portion of a genome.

[0058] In some cases, generating nucleic acid sequence data includes detecting low allele fractions by deep sequencing. In some cases, deep sequencing is performed by next generation sequencing. In some cases, deep sequencing is performed by avoiding error-prone regions. In some cases, error-prone regions may include regions near sequence duplications, regions with abnormally high or low %GC, regions near homopolymers, di- and trinucleotides, and other short repeats. In some cases, error-prone regions may include regions that lead to DNA sequencing errors (e.g., polymerase slippage in homopolymer sequences).

[0059] In some cases, generating the nucleic acid sequence data includes performing one or more sequencing reactions on one or more nucleic acid molecules in the sample. Certain embodiments may include performing 1 or more, 2 or more, 3 or more, 4 or more, 5 or more, 6 or more, 7 or more, 8 or more, 9 or more, 10 or more, 15 or more, 20 or more, 30 or more, 40 or more, 50 or more, 60 or more, 70 or more, 80 or more, 90 or more, 100 or more, 200 or more, 300 or more, 400 or more, 500 or more, 600 or more, 700 or more, 800 or more, 900 or more, or 1000 or more sequencing reactions on one or more nucleic acid molecules in the sample. The sequencing reactions may be performed simultaneously, sequentially, or a combination thereof. The sequencing reactions may include whole genome sequencing or exome sequencing. The sequencing reactions may include a Maxim-Gilbert system, a chain termination system, or a high throughput system. Alternatively or additionally, the sequencing reaction may include HeliScope™ single molecule sequencing, nanopore DNA sequencing, Massively Parallel Signature Sequencing (MPSS) by Lynx Therapeutics, 454 pyrosequencing, single molecule real-time (RNAP) sequencing, Illumina (Solexa) sequencing, SOLiD sequencing, Ion Torrent™ ion semiconductor sequencing, single molecule SMRT™ sequencing, polony sequencing, DNA nanoball sequencing, an approach by VisiGen Biotechnologies, or a combination thereof.Alternatively or additionally, the sequencing reaction may include one or more sequencing platforms, such as, but not limited to, Single Molecule Real Time (SMRT™) technology, such as the Genome Analyzer IIx, HiSeq, and MiSeq from Illumina, PacBio RS system from Pacific Biosciences (CA), and Solexa sequencer (Illumina), and True Single Molecule Sequencing (tSMS™) technology, such as the HeliScope™ sequencer from Helicos (Cambridge, MA). The sequencing reaction may also include an electron microscope or a chemically sensitive field effect transistor (chemFET) array. In some embodiments of the present disclosure, the sequencing reaction comprises capillary sequencing, next generation sequencing, Sanger sequencing, sequencing by synthesis, sequencing by ligation, sequencing by hybridization, single molecule sequencing, or a combination thereof. Sequencing by synthesis may comprise reversible terminator sequencing, processive single molecule sequencing, sequence flow sequencing, or a combination thereof. Sequence flow sequencing may comprise pyrosequencing, pH-mediated sequencing, semiconductor sequencing, or a combination thereof.

[0060] In some cases, generating the nucleic acid sequence data includes performing at least one long read sequencing reaction and at least one short read sequencing reaction. The long read sequencing reaction and / or the short read sequencing reaction may be performed on at least a portion of the subset of nucleic acid molecules. The long read sequencing reaction and / or the short read sequencing reaction may be performed on at least a portion of two or more subsets of nucleic acid molecules. Both the long read sequencing reaction and the short read sequencing reaction may be performed on at least a portion of one or more subsets of nucleic acid molecules.

[0061] The sequencing of one or more nucleic acid molecules or a subset thereof may be at least about 5, 10, 15, 20, 25, 30, 35, 40, 45, 50, 60, 70, 80, 90, 100, 200, 300, 400, 500, 600, 700, 800, 900, 1,000, 1500, 2,000, 2500, 3,000, 3500, 4,000, 4500, 5,000, 5500, 6,000, 6500, 7,000, 7500, 8,000, 9,000, 10,000, 11,000, 12,000, 13,000, 14,000, 15,000, 16,000, 17,000, 18,000, 19,000, 21,000, 22,000, 23,000, 24,000, 25,000, 26,000, 27,000, 28,000, 29,000, 30,000, 31,000, 32,000, 33,000, 34,000, 35,000, 36,000, 37,000, 38,000, 39,000, 40,000, 41,000, 42,000, 43,000, 44,000, 45,000, 46,000, In one embodiment, the sequence may include 00, 8500, 9,000, 10,000, 25,000, 50,000, 75,000, 100,000, 250,000, 500,000, 750,000, 10,000,000, 25,000,000, 50,000,000, 100,000,000, 250,000,000, 500,000,000, 750,000,000, 1,000,000,000 or more sequencing reads.

[0062] The sequencing reaction may comprise at least about 50, 60, 70, 80, 90, 100, 110, 120, 130, 140, 150, 160, 170, 180, 190, 200, 210, 220, 230, 240, 250, 260, 270, 280, 290, 300, 325, 350, 375, 400, 425, 450, 475, 500, 600, 700, 800, 900, 1,000, 1500, In some embodiments, the method may include sequencing 2,000, 2500, 3,000, 3500, 4,000, 4500, 5,000, 5500, 6,000, 6500, 7,000, 7500, 8,000, 8500, 9,000, 10,000, 20,000, 30,000, 40,000, 50,000, 60,000, 70,000, 80,000, 90,000, 100,000 or more bases or base pairs. The sequencing reaction may comprise at least about 50, 60, 70, 80, 90, 100, 110, 120, 130, 140, 150, 160, 170, 180, 190, 200, 210, 220, 230, 240, 250, 260, 270, 280, 290, 300, 325, 350, 375, 400, 425, 450, 475, 500, 600, 700, 800, 900, 1,000, 1500, 2, The method may include sequencing 000, 2500, 3,000, 3500, 4,000, 4500, 5,000, 5500, 6,000, 6500, 7,000, 7500, 8,000, 8500, 9,000, 10,000, 20,000, 30,000, 40,000, 50,000, 60,000, 70,000, 80,000, 90,000, 100,000 or more consecutive bases or base pairs.

[0063] In some cases, the sequencing technique generates at least 100 reads / run (reads / run), at least 200 reads / run, at least 300 reads / run, at least 400 reads / run, at least 500 reads / run, at least 600 reads / run, at least 700 reads / run, at least 800 reads / run, at least 900 reads / run, at least 1000 reads / run, at least 5,000 reads / run, at least 10,000 reads / run, at least 50,000 reads / run, at least 100,000 reads / run, at least 500,000 reads / run, or at least 1,000,000 reads / run. Alternatively, the sequencing technique generates at least 1,500,000 reads / run, at least 2,000,000 reads / run, at least 2,500,000 reads / run, at least 3,000,000 reads / run, at least 3,500,000 reads / run, at least 4,000,000 reads / run, at least 4,500,000 reads / run, or at least 5,000,000 reads / run.

[0064] In some cases, the sequencing technique generates at least about 30 base pairs, at least about 40 base pairs, at least about 50 base pairs, at least about 60 base pairs, at least about 70 base pairs, at least about 80 base pairs, at least about 90 base pairs, at least about 100 base pairs, at least about 110, at least about 120 base pairs per read, at least about 150 base pairs, at least about 200 base pairs, at least about 250 base pairs, at least about 300 base pairs, at least about 350 base pairs, at least about 400 base pairs, at least about 450 base pairs, at least about 500 base pairs, at least about 550 base pairs, at least about 600 base pairs, at least about 700 base pairs, at least about 800 base pairs, at least about 900 base pairs, or at least about 1,000 base pairs per read. Additionally or alternatively, the sequencing technique may generate long sequencing reads. In some cases, the sequencing technique provides a sequence of at least about 1,200 base pairs / read, at least about 1,500 base pairs / read, at least about 1,800 base pairs / read, at least about 2,000 base pairs / read, at least about 2,500 base pairs / read, at least about 3,000 base pairs / read, at least about 3,500 base pairs / read, at least about 4,000 base pairs / read, at least about 4,500 base pairs / read, at least about 5,000 base pairs / read, at least about 6,000 base pairs / read, In one embodiment, the method may generate at least about 10,000 base pairs / read, at least about 7,000 base pairs / read, at least about 8,000 base pairs / read, at least about 9,000 base pairs / read, at least about 10,000 base pairs / read, 20,000 base pairs / read, 30,000 base pairs / read, 40,000 base pairs / read, 50,000 base pairs / read, 60,000 base pairs / read, 70,000 base pairs / read, 80,000 base pairs / read, 90,000 base pairs / read, or 100,000 base pairs / read.

[0065] High-throughput sequencing systems may allow detection of sequenced nucleotides immediately or as they are incorporated into the growing strand, i.e., detection of sequences in real time or substantially real time. In some cases, high-throughput sequencing generates at least 1,000, at least 5,000, at least 10,000, at least 20,000, at least 30,000, at least 40,000, at least 50,000, at least 100,000, or at least 500,000 sequence reads / hour, with each read being at least 50, at least 60, at least 70, at least 80, at least 90, at least 100, at least 120, at least 150, at least 200, at least 250, at least 300, at least 350, at least 400, at least 450, or at least 500 bases / read. Sequencing may be performed using a nucleic acid as described herein, such as genomic DNA, cDNA derived from an RNA transcript, or RNA as a template.

[0066] C. Identification of Nucleotide Sequence Variants To estimate the tumor purity of the biological sample, the B allele frequency (BAF) of the biological sample can be determined. To determine the BAF, a set of genomic regions can be identified, and each of the set of genomic regions can include nucleotide sequence variants associated with the corresponding genomic region of the reference genome. To determine the set of genomic regions, the nucleic acid sequence data can be aligned to the reference genome. For example, the FASTQ file corresponding to the nucleic acid sequence data can be aligned to the reference genome to generate one or more BAM files. The one or more BAM files can be processed by another module (e.g., GATK HaplotypeCaller) to identify the set of genomic regions. In some cases, a VCF file is generated that includes the set of genomic regions. Additionally or alternatively, the VCF file can be obtained for the matched paired cell line. The VCF file can identify the set of genomic regions that have nucleotide sequence variants based on the comparison of the nucleic acid sequence data to the reference genome.

[0067] Certain embodiments may include a nucleic acid molecule comprising one or more genomic regions. Certain embodiments may include a nucleic acid molecule comprising one or more sets of genomic regions. The one or more genomic regions may comprise one or more genomic region features. The genomic region features may comprise an entire genome or a portion thereof. The genomic region features may comprise an entire exome or a portion thereof. The genomic region features may comprise one or more gene sets. The genomic region features may comprise one or more genes. The genomic region features may comprise one or more genes. The genomic region features may comprise one or more sets of regulatory elements. The genomic region features may comprise one or more regulatory elements.

[0068] The genomic region features may include a set of polymorphisms. The genomic region features may include one or more polymorphisms. The genomic region features may be related to the GC content, complexity, and / or mappability of one or more nucleic acid molecules. The genomic region features may include one or more simple tandem repeats (STRs), unstable expansion repeats, segmental duplications, single and paired read degenerate mapping scores, GRCh37 patches, or combinations thereof. The genomic region features may include one or more low average coverage regions from whole genome sequencing (WGS), zero average coverage regions from WGS, validated compaction, or combinations thereof. The genomic region features may include one or more alternative or non-reference sequences. The genomic region features may include one or more gene phasing and reconstituted genes. In some aspects of the present disclosure, the one or more genomic region features are not mutually exclusive. For example, a feature of a genomic region that includes an entire genome or a portion thereof may overlap with a feature of an additional genomic region, such as an entire exome or a portion thereof, one or more genes, one or more regulatory elements, etc. Alternatively, one or more features of a genomic region are mutually exclusive. For example, a genomic region that includes a non-coding portion of an entire genome may not overlap with a feature of a genomic region, such as an exome or a portion thereof, or a coding portion of a gene. Alternatively or additionally, one or more features of a genomic region are partially exclusive or partially inclusive. For example, a genomic region that includes an entire exon or a portion thereof may partially overlap with a genomic region that includes an exon portion of a gene. However, a genomic region that includes an entire exome or a portion thereof may not overlap with a genomic region that includes an intron portion of a gene. Thus, a feature of a genomic region that includes a gene or a portion thereof may be partially exclusive and / or partially inclusive of a feature of a genomic region that includes an entire exome or a portion thereof.

[0069] Certain embodiments may include a nucleic acid sample or molecule comprising one or more genomic regions, at least one of which comprises a genomic region characteristic of an entire genome or a portion thereof. The entire genome or a portion thereof may comprise one or more coding portions of the genome, one or more non-coding portions of the genome, or a combination thereof. The coding portion of the genome may comprise one or more coding portions of genes encoding one or more proteins. The one or more coding portions of the genome may comprise an entire exome or a portion thereof. Alternatively or additionally, the one or more coding portions of the genome may comprise one or more exons.

[0070] The one or more non-coding portions of the genome may include one or more non-coding molecules or parts thereof. The non-coding molecules may include one or more non-coding RNAs, one or more regulatory elements, one or more introns, one or more pseudogenes, one or more repeat sequences, one or more transposons, one or more viral elements, one or more telomeres, parts thereof, or combinations thereof. The non-coding RNA may be a functional RNA molecule that is not translated into a protein. Examples of non-coding RNA include, but are not limited to, ribosomal RNA, transfer RNA, PIWI-interacting RNA, microRNA, siRNA, shRNA, snoRNA, sncRNA, and lncRNA. A pseudogene may be associated with a known gene and is typically no longer expressed. The repeat sequence may include one or more tandem repeats, one or more interspersed repeats, or a combination thereof. The tandem repeats may include one or more satellite DNA, one or more minisatellites, one or more microsatellites, or a combination thereof.

[0071] Interspersed repeats may include one or more transposons. Transposons may be mobile genetic elements. Mobile genetic elements are often variable in location within the genome. Transposons can be classified as class I transposons (class I TEs) or class II transposons (class II TEs). Class I TEs (e.g., retrotransposons) often copy themselves in two steps, first from DNA to RNA by transcription, and then from RNA back to DNA by reverse transcription. The DNA copy can then be inserted into a new location in the genome. Class I TEs may include one or more long terminal repeats (LTRs), one or more long interspersed nuclear elements (LINEs), one or more short interspersed nuclear elements (SINEs), or a combination thereof. Examples of LTRs include, but are not limited to, human endogenous retroviruses (HERVs), medium reiterated repeats 4 (MER4), and retrotransposons. Examples of LINEs include, but are not limited to, LINE1 and LINE2. SINEs may include one or more Alu sequences, one or more mammalian-wide interspersed repeats (MIRs), or combinations thereof. Class II TEs (e.g., DNA transposons) often do not include an RNA intermediate. DNA transposons are often excised from one site and inserted at another site in the genome. Alternatively, DNA transposons are replicated and inserted at a new location in the genome. Examples of DNA transposons include, but are not limited to, MER1, MER2, and mariner. Viral elements may include one or more endogenous retroviral sequences. Telomeres are often regions of repetitive DNA at the ends of chromosomes.

[0072] Certain embodiments may include a subset of nucleic acid samples or nucleic acid molecules comprising one or more genomic regions, at least one of which comprises a genomic region characteristic of an entire exome or a portion thereof. An exome is often a portion of a genome formed by exons. An exome may be formed by untranslated regions (UTRs), splice sites, and / or intronic regions. An entire exome or a portion thereof may include one or more exons of a protein-coding gene. An entire exome or a portion thereof may include one or more untranslated regions (UTRs), splice sites, and introns.

[0073] Certain embodiments may include a nucleic acid sample or molecule comprising one or more genomic regions, at least one of which comprises a genomic region characteristic of a gene or a portion thereof. Typically, a gene comprises a stretch of nucleic acid that encodes a polypeptide or functional RNA. A gene may comprise one or more exons, one or more introns, one or more untranslated regions (UTRs), or a combination thereof. Exons often code for sections of a gene, which are transcribed into precursor mRNA sequences and are found in the final mature RNA product of the gene. Introns often are non-coding sections of a gene, which are transcribed into precursor mRNA sequences and are removed by RNA splicing. UTRs may refer to the sections on either side of a coding sequence of an mRNA strand. A UTR that is 5' to a coding sequence may be referred to as a 5'UTR (or leader sequence). A UTR that is located 3' to a coding sequence may be referred to as a 3'UTR (or trailer sequence). A UTR may comprise one or more elements for controlling gene expression. Elements such as regulatory elements may be located in the 5'UTR. Regulatory sequences such as polyadenylation signals, protein binding sites, and miRNA binding sites may be located in the 3'UTR. Protein binding sites located in the 3'UTR include, but are not limited to, selenocysteine ​​insertion sequence (SECIS) elements and AU-rich elements (AREs). SECIS elements may instruct ribosomes to translate the codon UGA as selenocysteine ​​rather than a stop codon. AREs are often stretches that consist mainly of adenine and uracil nucleotides and may affect mRNA stability.

[0074] Certain embodiments may include a subset of nucleic acid samples or molecules comprising one or more genomic regions, at least one of which comprises a genomic region feature comprising a set of genes. The set of genes includes, but is not limited to, Mendel DB genes, Human Gene Mutation Database (HGMD) genes, Cancer Gene Census genes, Online Mendelian Inheritance in Man (OMIM) genes, HGMD Mendelian genes, and Human Leukocyte Antigen (HLA) genes. The set of genes may have one or more known Mendelian traits, one or more known disease traits, one or more known drug traits, one or more known biomedically interpretable variants, or a combination thereof. The Mendelian traits may be controlled by a single genetic locus and exhibit a Mendelian inheritance pattern. A set of genes with known Mendelian traits may include one or more genes encoding Mendelian traits including, but not limited to, taste for phenylthiocarbamide (dominant), smell for hydrogen cyanide (bitter almond-like) (recessive), albinism (recessive), brachydactyly (short fingers and toes), and wet (dominant) or dry (recessive) ear wax. A disease trait may cause or increase the risk of a disease and may be inherited in a Mendelian or complex pattern. A set of genes with known disease traits may include one or more genes encoding disease traits including, but not limited to, cystic fibrosis, hemophilia, and Lynch syndrome.

[0075] A drug trait may alter the metabolism, optimal dose, adverse reactions, and side effects of one or more drugs or drug families. The set of genes with known drug traits may include one or more genes encoding drug traits, including but not limited to CYP2D6, UGT1A1, and ADRB1. The biomedically interpretable variant may be a polymorphism of a gene associated with a disease or indication. The set of genes with known biomedically interpretable variants may include one or more genes encoding biomedically interpretable variants, including but not limited to cystic fibrosis (CF) mutations, muscular dystrophy mutations, p53 mutations, Rb mutations, cell cycle regulators, receptors, and kinases. Alternatively or additionally, the set of genes with known biomedically interpretable variants may include one or more genes associated with Huntington's disease, cancer, cystic fibrosis, muscular dystrophy (e.g., Duchenne muscular dystrophy).

[0076] Certain embodiments may include a nucleic acid sample or molecule comprising one or more genomic regions, at least one of which comprises a genomic region characteristic of a regulatory element or a portion thereof. Regulatory elements may be cis-regulatory elements or trans-regulatory elements. Cis-regulatory elements may be sequences that control the transcription of nearby genes. Cis-regulatory elements may be located within 5' or 3' untranslated regions (UTRs) or introns. Trans-regulatory elements may control the transcription of distant genes. Regulatory elements may include one or more promoters, one or more enhancers, or a combination thereof. Promoters may enhance the transcription of a particular gene and may be found upstream of a coding region. Enhancers may have a distant effect on the transcription level of a gene.

[0077] Certain embodiments may include a subset of nucleic acid samples or molecules that include one or more genomic regions, at least one of which includes a genomic region feature that includes a polymorphism or a portion thereof. Generally, a polymorphism refers to a mutation of a genotype. A polymorphism may be a germline variant or a somatic variant. A polymorphism may include one or more base changes, insertions, repeats, or deletions of one or more bases. Copy number polymorphisms (CNVs), transversions, and other rearrangements are also forms of genetic variation. Polymorphic markers include restriction fragment length polymorphisms, variable number of tandem repeats (VNTRs), hypervariable regions, minisatellites, dinucleotide repeats, trinucleotide repeats, tetranucleotide repeats, simple sequence repeats, and insertion elements such as Alu. The allelic form that occurs most frequently in a selected population may be referred to as the wild type form. Diploid organisms may be homozygous or heterozygous for an allelic form. There are two forms of diallelic polymorphisms. There are three forms of triallelic polymorphisms. Single nucleotide polymorphisms (SNPs) are one form of polymorphism. In some aspects of the present disclosure, the one or more polymorphisms include one or more single nucleotide changes, indels, small insertions, small deletions, structural variant junctions, variable length tandem repeats, flanking sequences, or combinations thereof. The one or more polymorphisms may be located in coding and / or non-coding regions. The one or more polymorphisms may be located in, around, or near genes, exons, introns, splice sites, untranslated regions, or combinations thereof. The one or more polymorphisms may span at least a portion of a gene, exon, intron, untranslated region.

[0078] Certain embodiments may include a nucleic acid sample or molecule comprising one or more genomic regions, at least one of which comprises a genomic region feature comprising one or more simple tandem repeats (STRs), unstable expansion repeats, segmental duplications, single and paired read degenerate mapping scores, GRCh37 patches, or combinations thereof. The one or more STRs may comprise one or more homopolymers, one or more dinucleotide repeats, one or more trinucleotide repeats, or combinations thereof. The one or more homopolymers may be about 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20 or more bases or base pairs. The dinucleotide and / or trinucleotide repeats can be about 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 35, 40, 45, 50 or more bases or base pairs. Single and paired read degenerate mapping scores can be based on or derived from 100mer alignability by GEM in ENCODE / CRG (Guigo), 75mer alignability by GEM in ENCODE / CRG (Guigo), 100 base pair boxcar average of signal mappability, trajectory of paired read scores and maximum possible pairs, or combinations thereof.

[0079] Genomic region features may include one or more low average coverage regions from whole genome sequencing (WGS), zero average coverage regions from WGS, validated compactions, or combinations thereof. WGS low average coverage regions may include regions generated by Illumina® V3 chemistry, regions below the first percentile of a Poisson distribution based on average coverage, or combinations thereof. WGS zero average coverage regions may include regions generated by Illumina® V3 chemistry. Validated compactions may include regions with high mapped depth, regions with two or more observed haplotypes, regions expected to be missing repeats in the reference, or combinations thereof. Genomic region features may include one or more alternative or non-reference sequences. One or more alternative or non-reference sequences may include known structural variant junctions, known insertions, known deletions, alternative haplotypes, or combinations thereof. Genomic region features may include one or more gene phasing and rearrangement genes. Examples of phasing and rearrangement genes include, but are not limited to, one or more of major histocompatibility complex, blood group, and amylase gene family. One or more major histocompatibility complex may include one or more of HLA class I, HLA class II, or a combination thereof. One or more HLA class I may include HLA-A, HLA-B, HLA-C, or a combination thereof. One or more HLA class II may include HLA-DP, HLA-DM, HLA-DOA, HLA-DOB, HLA-DQ, HLA-DR, or a combination thereof. Blood group genes may include ABO, RHD, RHCE, or a combination thereof.

[0080] Certain embodiments may include a nucleic acid sample or molecule comprising one or more genomic regions, at least one of which comprises a genomic region feature associated with the GC content of one or more nucleic acid molecules. The GC content may refer to the GC content of the nucleic acid molecule. Alternatively, the GC content may refer to the GC content of one or more nucleic acid molecules, and may be referred to as the average GC content. As used herein, the terms "GC content" and "average GC content" may be used interchangeably. The GC content of the genomic region may be a high GC content. Typically, a high GC content refers to a GC content of about 65%, 70%, 75%, 80%, 85%, 90%, 95%, 97%, or more. In some aspects of the present disclosure, a high GC content may refer to a GC content of about 70% or more. The GC content of the genomic region may be a low GC content. Typically, low GC content refers to a GC content of about 65%, 60%, 55%, 50%, 45%, 40%, 35%, 30%, 25%, 20%, 15%, 10%, 5%, 2% or less.

[0081] Certain embodiments may include a nucleic acid sample or molecule comprising one or more genomic regions, at least one of which comprises a genomic region feature associated with the complexity of one or more nucleic acid molecules. The complexity of a nucleic acid molecule may refer to the randomness of the nucleotide sequence. Low complexity may refer to the pattern, repetition, and / or lack of one or more nucleotide species within a sequence.

[0082] Certain embodiments may include a nucleic acid sample or molecule that comprises one or more genomic regions, at least one of which comprises a genomic region feature that is related to the mappability of one or more nucleic acid molecules. The mappability of a nucleic acid molecule may refer to the uniqueness of its alignment to a reference sequence. A nucleic acid molecule that has low mappability may be poorly aligned with a reference sequence.

[0083] D. Calculation of BAF distribution For each genomic region of the set of genomic regions, the B allele frequency (BAF) can be determined. BAF refers to a normalized measure of the allele intensity ratio of two alleles (A and B). A BAF of 100 or 0 indicates the complete absence of one of the two alleles (e.g., AA or BB). For example, if both chromosomes of a sample contain a nucleotide sequence variant related to the reference genome, the BAF of a particular genomic region may be 100. Thus, the variant allele should be present in nearly 100% of the reads of that sample. In some cases, the BAF distribution is normalized to generate a density plot such that the sum of the areas under the curve is 1. Specifically, a histogram of the BAFs of all variants is created, the area of ​​this histogram is calculated, and each bin of the histogram is divided by the calculated area.

[0084] For heterozygous alleles in a sample, the variant is present in a subset of sequence reads. Thus, a BAF of 50 indicates that both alleles (e.g., AB) are equally present. In normal samples without copy number variation, a BAF of 0, 50, or 100 is expected. However, in tumor samples, mutations can alter the BAF value. For example, a copy number event that duplicates the B allele may lead to a B allele frequency of -67%. In another example, if only one chromosome contains the variant, the variant allele should be present in nearly 50% of the reads for that sample. Random variations inherent in the data generation process introduce noise, causing the observed BAF measurements to deviate from the ideal value for a particular copy number in both tumor and normal samples.

[0085] In some cases, a subset of genomic regions corresponding to heterozygous sites in a biological sample (e.g., an exome sample) is identified, and the allele frequency of each of the subset of genomic regions is quantified. Thus, the VCF file can be filtered to obtain the reference and alternative read depths of all heterozygous sites that have nucleotide sequence variants. Such information can be used to calculate the BAF of each genomic region of the subset of genomic regions that have heterozygosity.

[0086] Based on the BAF determined for each set of genomic regions, a normalized BAF distribution can be calculated. The BAF distribution can indicate absolute amounts, percentages, and / or normalized amounts corresponding to the BAF of each genomic region of the set of genomic regions. In some cases, the normalized BAF distribution is determined from the BAFs corresponding to a subset of genomic regions (e.g., heterozygous sites across exomes). Since BAF values ​​are typically 0-100, the normalized BAF distribution can include up to 101 BAF values, with each value indicating a corresponding normalized frequency. Additionally or alternatively, the normalized BAF distribution can be modified such that the corresponding BAF values ​​can be populated into a predefined number of bins. For example, each bin can correspond to a non-overlapping range of BAF values ​​(e.g., 1-9, 10-19), and the BAF values ​​can be assigned to the corresponding bins. The BAF distribution can then be determined based on the values ​​corresponding to each predefined bin.

[0087] 7A-7E provide examples of plotted BAF distributions according to some embodiments. In both figures, FIG. 7A and FIG. 7B, the x-axis represents BAF values ​​ranging from 0% (homozygous for the A allele) to 100% (homozygous for the B allele). As mentioned above, a BAF of 50 indicates heterozygosity, which corresponds to both alleles (e.g., AB) being equally present. In normal samples, BAFs of 0, 50, or 100 may be expected. However, in tumor samples, mutations may alter the BAF values. For example, a copy number event that duplicates the B allele may lead to a B allele frequency of -67%. The y-axis represents normalized frequency values ​​corresponding to each BAF value on the x-axis. The normalized frequency values ​​may identify several genomic regions that correspond to specific BAF values. For example, the normalized frequency of a BAF of 50% may be greater than 4.

[0088] FIG. 7A is a histogram of B allele frequency from heterozygous sites in chromosomes of pure normal samples. The histogram shows a normal distribution, with most observed B allele frequencies dropping close to 50%. FIG. 7B is a histogram of B allele frequency from heterozygous sites in pure tumor history samples. The histogram shows an altered distribution, with more observed B allele frequencies dropping further from 50%. As shown in FIG. 7B, it can be seen that an increase in the number of somatic mutations may have altered the BAF frequency from 50%.

[0089] FIG. 7C is a heat map, where each row represents the BAF distribution from chromosomes of pure normal samples. The heat map shows a normal distribution, with most observed B allele frequencies dropping close to 50%. FIG. 7E is a heat map, where each row represents the BAF distribution from chromosomes of pure tumor samples. The heat map shows an altered distribution, with more observed B allele frequencies dropping further from 50%. FIG. 7D is a heat map from a sample that is 50% normal and 50% tumor. The heat map shows an intermediate distribution between the distributions shown in FIG. 7C and FIG. 7E.

[0090] E. Classification of pure tumor and pure normal samples using BAF In some cases, the BAF signature of a biological sample is used to classify the biological sample as normal or tumorous. For example, a normalized BAF distribution can be calculated at heterozygous sites across the exomes of multiple pure tumor samples and pure normal samples. Pure tumor samples and pure normal samples can be classified using logistic regression using the whole exome BAF signature.

[0091] Figure 8 shows a plot 800 of the first two main components (PC1 and PC2) for classifying biological samples using B allele frequency features, according to some embodiments. In Figure 8, the first two main components show near-linear separability of tumor and normal samples, indicating that BAF features may be suitable for estimating the tumor content of samples. Furthermore, it is demonstrated that pure tumor and pure normal samples can be classified using logistic regression using whole-exome BAF features.

[0092] F. Estimation of tumor purity based on BAF distribution A trained machine learning model may be used to process the BAF distribution of the biological sample to estimate a metric that identifies the tumor purity of the biological sample. In some embodiments, the trained machine learning model includes a fully connected neural network. The fully connected neural network may include a fully connected layer with a Rectified Linear Unit (ReLU) activation function. In some embodiments, the output activation function of the fully connected neural network is a sigmoid function. A loss function of the fully connected neural network may be configured to generate a mean squared error (MSE). In some embodiments, the fully connected neural network is tuned by a hyperparameter search using random sampling with a linear search of layers, a linear search of size, a logarithmic search of learning rate, or a combination thereof.

[0093] In some cases, the trained machine learning model includes a one-dimensional convolutional neural network that can use the B allele frequency distribution as an input and can be configured to be encoded to an input size of height 25, width 1, and depth 100. In some embodiments, each layer of the one-dimensional convolutional neural network performs a 1×1 convolution followed by a ReLU activation function.

[0094] In some embodiments, the trained machine learning model includes a two-dimensional convolutional neural network. In some embodiments, the B allele frequency distribution of the two-dimensional convolutional neural network is used as an input and encoded into an input size defined as height 25, width 100, and depth 1. In some embodiments, each convolutional layer of the two-dimensional convolutional neural network is followed by a ReLU layer. The output of the two-dimensional convolutional neural network may be a densely connected layer with a sigmoid activation function. Additionally, the hyperparameters of the two-dimensional convolutional neural network may be tuned by adjusting the number of layers, the filter size, the number of filters, or a combination thereof.

[0095] An estimated metric identifying tumor purity may be output. For example, a report including the estimated metric may be output. In some embodiments, the report includes information identifying a B allele frequency distribution. The report may also include information identifying at least one diagnostic marker and / or at least one prognostic marker. In some embodiments, the report includes information identifying a predicted somatic variant. The report may also include a treatment recommendation. In some embodiments, the treatment recommendation includes a recommendation to administer a treatment to the human subject. The treatment recommendation may include a recommendation not to administer a treatment to the human subject.

[0096] IV. Example Process for Estimating Tumor Purity from a Single Sample FIG. 9 includes a flowchart 900 illustrating an example of a method for estimating tumor purity of a biological sample according to certain embodiments. The operations described in the flowchart 900 may be performed by a computer system implementing a trained machine learning model, such as, for example, a trained one-dimensional or two-dimensional convolutional neural network. Although the flowchart 900 may describe the operations as a sequential process, in various embodiments, many of the operations may be performed in parallel or simultaneously. Also, the order of the operations may be changed. The operations may include additional steps not shown. Furthermore, embodiments of the method may be implemented by hardware, software, firmware, middleware, microcode, hardware description languages, or any combination thereof. When implemented in software, firmware, middleware, or microcode, the program code or code segments for performing the associated tasks may be stored in a computer-readable medium, such as a storage medium.

[0097] At operation 910, the computer system obtains nucleic acid sequence data representing a plurality of nucleic acid molecules of a biological sample of a subject. The nucleic acid sequence data may be generated by sequencing a plurality of nucleic acid molecules of a tumor sample. In some cases, the plurality of nucleic acid molecules are isolated prior to sequencing. The nucleic acid sequence data may correspond to whole exome sequence data. Alternatively or additionally, the nucleic acid sequence data is whole genome sequencing data.

[0098] In operation 920, the computer system aligns the nucleic acid sequence data to a reference genome. For example, a FASTQ file corresponding to the nucleic acid sequence data can be aligned to a reference genome to generate one or more BAM files.

[0099] At operation 930, the computer system identifies a set of genomic regions based on the aligned nucleic acid sequence data. Each genomic region of the set of genomic regions may include one or more nucleotide sequence variants associated with a corresponding genomic region of the reference genome. In some cases, the computer system identifies one or more candidate nucleotide sequence variants in the nucleic acid sequence data and calculates a reference read depth and an alternative read depth for each of the one or more candidate nucleotide sequence variants.

[0100] In operation 940, the computer system measures the BAF of each genomic region of the set of genomic regions. BAF refers to a normalized measurement of the allele intensity ratio of two alleles (A and B). In some cases, a BAF of 100 or 0 indicates the complete absence of one of the two alleles (e.g., AA or BB), and a BAF of 50 indicates the equal presence of both alleles (e.g., AB). In normal samples, a BAF of 0, 50, or 100 may be expected. However, in tumor samples, BAF values ​​may change due to mutations.

[0101] In operation 950, the computer system determines a BAF distribution of the biological sample based on the BAFs of the set of genomic regions. Optionally, the B allele frequencies are normalized.

[0102] At operation 960, the computer system processes the B allele frequency distribution using a trained machine learning model to estimate a metric identifying tumor purity of the biological sample. In some cases, the trained machine learning model was trained with a training data set generated from nucleic acid sequence data from one or more tumor cells diluted with normal cells. The mean absolute error of the trained machine learning model may be less than about 0.2.

[0103] At operation 970, the computer system outputs a metric. In some cases, the computer system outputs a report including an estimated metric identifying tumor purity. In addition to the estimated metric, the report may include a measured B allele frequency distribution. The report may also include other types of information, including, but not limited to, at least one biomarker, at least one prognostic marker, predicted somatic variants, and a treatment recommendation (e.g., a recommendation as to whether a treatment should be administered to the subject).

[0104] V. Additional Considerations A. Probing Techniques Certain embodiments may include one or more labels. The one or more labels may be attached to one or more capture probes, nucleic acid molecules, beads, primers, or combinations thereof. Examples of labels include, but are not limited to, detectable labels such as radioisotopes, fluorophores, chemiluminophores, chromophores, lumiphores, enzymes, colloidal particles, fluorescent microparticles, quantum dots, and antigens, antibodies, haptens, avidin / streptavidin, biotin, haptens, enzyme cofactors / substrates, one or more members of a quenching system, chromogens, haptens, magnetic particles, materials exhibiting nonlinear optics, semiconductor nanocrystals, metal nanoparticles, enzymes, aptamers, and one or more members of a binding pair.

[0105] Certain embodiments may include one or more capture probes, multiple capture probes, or one or more capture probe sets. Typically, the capture probe comprises a nucleic acid binding site. The capture probe may further comprise one or more linkers. The capture probe may further comprise one or more labels. The one or more linkers may attach one or more labels to the nucleic acid binding site.

[0106] The capture probe may hybridize to one or more nucleic acid molecules in the sample. The capture probe may hybridize to one or more genomic regions. The capture probe may hybridize to one or more genomic regions within, surrounding, adjacent to, or spanning one or more genes, exons, introns, UTRs, or combinations thereof. The capture probe may hybridize to one or more genomic regions spanning one or more genes, exons, introns, UTRs, or combinations thereof. The capture probe may hybridize to one or more known indels. The capture probe may hybridize to one or more known structural variants.

[0107] Certain embodiments may include 1 or more, 2 or more, 3 or more, 4 or more, 5 or more, 6 or more, 7 or more, 8 or more, 9 or more, 10 or more, 20 or more, 30 or more, 40 or more, 50 or more, 60 or more, 70 or more, 80 or more, 90 or more, 100 or more, 125 or more, 150 or more, 175 or more, 200 or more, 250 or more, 300 or more, 350 or more, 400 or more, 500 or more, 600 or more, 700 or more, 800 or more, 900 or more, 1000 or more capture probes or capture probe sets, which may be different, similar, identical, or a combination thereof.

[0108] The one or more capture probes may comprise a nucleic acid binding site that hybridizes to at least a portion of one or more nucleic acid molecules or variants or derivatives thereof in a sample or subset of nucleic acid molecules. The capture probes may comprise a nucleic acid binding site that hybridizes to one or more genomic regions. The capture probes may hybridize to different, similar, and / or identical genomic regions. The one or more capture probes may be at least about 50%, 55%, 60%, 65%, 70%, 75%, 80%, 85%, 90%, 95%, 97%, 99% or more complementary to the one or more nucleic acid molecules or variants or derivatives thereof.

[0109] The capture probe may comprise one or more nucleotides. The capture probe may comprise 1 or more, 2 or more, 3 or more, 4 or more, 5 or more, 6 or more, 7 or more, 8 or more, 9 or more, 10 or more, 20 or more, 30 or more, 40 or more, 50 or more, 60 or more, 70 or more, 80 or more, 90 or more, 100 or more, 125 or more, 150 or more, 175 or more, 200 or more, 250 or more, 300 or more, 350 or more, 400 or more, 500 or more, 600 or more, 700 or more, 800 or more, 900 or more, or 1000 or more nucleotides. The capture probe may comprise about 100 nucleotides. The capture probe may comprise about 10 to about 500 nucleotides, about 20 to about 450 nucleotides, about 30 to about 400 nucleotides, about 40 to about 350 nucleotides, about 50 to about 300 nucleotides, about 60 to about 250 nucleotides, about 70 to about 200 nucleotides, or about 80 to about 150 nucleotides. In some embodiments of the present disclosure, the capture probe comprises about 80 nucleotides to about 100 nucleotides.

[0110] The plurality or capture probe set may include two or more capture probes having identical, similar, and / or different nucleic acid binding site sequences, linkers, and / or labels. For example, two or more capture probes include identical nucleic acid binding sites. In another example, two or more capture probes include similar nucleic acid binding sites. In another example, two or more capture probes include different nucleic acid binding sites. The two or more capture probes may further include one or more linkers. The two or more capture probes may further include different linkers. The two or more capture probes may further include similar linkers. The two or more capture probes may further include identical linkers. The two or more capture probes may further include one or more labels. The two or more capture probes may further include different labels. The two or more capture probes may further include similar labels. The two or more capture probes may further include identical labels.

[0111] B. Assays and Amplification Techniques Certain embodiments may include performing one or more assays on a sample that includes one or more nucleic acid molecules. Generating two or more subsets of nucleic acid molecules may include performing one or more assays. The assays may be performed on a subset of nucleic acid molecules from the sample. The assays may be performed on one or more nucleic acid molecules from the sample. The assays may be performed on at least a portion of the subset of nucleic acid molecules. The assays may include one or more techniques, reagents, capture probes, primers, labels, and / or components for detection, quantification, and / or analysis of one or more nucleic acid molecules.

[0112] An assay may include, but is not limited to, sequencing, amplification, hybridization, enrichment, isolation, elution, fragmentation, detection, quantification of one or more nucleic acid molecules. An assay may include a method for preparing one or more nucleic acid molecules.

[0113] Certain embodiments may include performing one or more amplification reactions on one or more nucleic acid molecules in a sample. The term "amplification" refers to any process that generates at least one copy of a nucleic acid molecule. The terms "amplicon" and "amplified nucleic acid molecule" refer to copies of a nucleic acid molecule and may be used interchangeably. The amplification reaction may include a PCR-based method, a non-PCR-based method, or a combination thereof. Examples of non-PCR-based methods include, but are not limited to, multiple displacement amplification (MDA), transcription-mediated amplification (TMA), nucleic acid sequence-based amplification (NASBA), strand displacement amplification (SDA), real-time SDA, rolling circle amplification, or circle-to-circle amplification. PCR-based methods include, but are not limited to, PCR, HD-PCR, next-generation PCR, digital RTA, or any combination thereof. Additional PCR methods include, but are not limited to, linear amplification, allele-specific PCR, Alu PCR, assembly PCR, asymmetric PCR, droplet PCR, emulsion PCR, helicase-dependent amplification HDA, hot-start PCR, inverse PCR, linear-post-exponential (LATE) PCR, long PCR, multiplex PCR, nested PCR, hemi-nested PCR, quantitative PCR, RT-PCR, real-time PCR, single-cell PCR, and touchdown PCR.

[0114] Certain embodiments may include performing one or more hybridizations to one or more nucleic acid molecules in a sample. The hybridization reaction may include hybridization of one or more capture probes to one or more nucleic acid molecules in a sample or a subset of nucleic acid molecules. The hybridization reaction may include hybridizing one or more capture probe sets to one or more nucleic acid molecules in a sample or a subset of nucleic acid molecules. The hybridization reaction may include one or more hybridization arrays, multiplex hybridization reactions, hybridization chain reactions, isothermal hybridization reactions, nucleic acid hybridization reactions, or combinations thereof. The one or more hybridization arrays may include hybridization array genotyping, hybridization array proportional sensing, DNA hybridization arrays, macroarrays, microarrays, high density oligonucleotide arrays, genomic hybridization arrays, comparative hybridization arrays, or combinations thereof. A hybridization reaction may include one or more capture probes, one or more beads, one or more labels, one or more subsets of nucleic acid molecules, one or more nucleic acid samples, one or more reagents, one or more wash buffers, one or more elution buffers, one or more hybridization buffers, one or more hybridization chambers, one or more incubators, one or more separators, or a combination thereof.

[0115] Certain embodiments may include performing one or more enrichment reactions on one or more nucleic acid molecules in a sample. The enrichment reaction may include contacting the sample with one or more beads or bead sets. The enrichment reaction may include differential amplification of two or more subsets of nucleic acid molecules based on features of one or more genomic regions. For example, the enrichment reaction includes differential amplification of two or more subsets of nucleic acid molecules based on GC content. Alternatively or additionally, the enrichment reaction includes differential amplification of two or more subsets of nucleic acid molecules based on methylation status. The enrichment reaction may include one or more hybridization reactions. The enrichment reaction includes isolation and / or purification of one or more hybridized nucleic acid molecules, one or more bead-bound nucleic acid molecules, one or more free nucleic acid molecules (e.g., nucleic acid molecules without capture probes, bead-unbound nucleic acid molecules), one or more labeled nucleic acid molecules, one or more unlabeled nucleic acid molecules, one or more amplicons, one or more unamplified nucleic acid molecules, or combinations thereof. Additionally or alternatively, the enrichment reaction may include enrichment of one or more cell types in the sample. The one or more cell types may be enriched by flow cytometry.

[0116] The enrichment reaction or reactions may produce one or more enriched nucleic acid molecules. The enriched nucleic acid molecules may include nucleic acid molecules or variants or derivatives thereof. For example, the enriched nucleic acid molecules include one or more hybridized nucleic acid molecules, one or more bead-bound nucleic acid molecules, one or more free nucleic acid molecules (e.g., nucleic acid molecules without capture probes, bead-unbound nucleic acid molecules), one or more labeled nucleic acid molecules, one or more unlabeled nucleic acid molecules, one or more amplicons, one or more unamplified nucleic acid molecules, or combinations thereof. The enriched nucleic acid molecules may be distinguished from non-enriched nucleic acid molecules by GC content, molecular size, genomic region, genomic region features, or combinations thereof. The enriched nucleic acid molecules may originate from one or more assays, supernatants, eluates, or combinations thereof. The enriched nucleic acid molecules may differ from non-enriched nucleic acid molecules by average size, average GC content, genomic region, or combinations thereof.

[0117] Certain embodiments may include performing one or more isolation or purification reactions on one or more nucleic acid molecules in a sample. The isolation or purification reaction may include contacting the sample with one or more beads or bead sets. The isolation or purification reaction may include one or more hybridization reactions, enrichment reactions, amplification reactions, sequencing reactions, or combinations thereof. The isolation or purification reaction may include the use of one or more separators. The one or more separators may include a magnetic separator. The isolation or purification reaction may include separating nucleic acid molecules bound to the beads from nucleic acid molecules not bound to the beads. The isolation or purification reaction may include separating nucleic acid molecules hybridized with a capture probe from nucleic acid molecules not hybridized with a capture probe. The isolation or purification reaction may include separating a first subset of nucleic acid molecules from a second subset of nucleic acid molecules, the first subset of nucleic acid molecules differing from the second subset of nucleic acid molecules in average size, average GC content, genomic region, or combinations thereof.

[0118] Certain embodiments may include performing one or more elution reactions on one or more nucleic acid molecules in a sample. The elution reaction may include contacting the sample with one or more beads or bead sets. The elution reaction may include separating nucleic acid molecules bound to the beads from nucleic acid molecules not bound to the beads. The elution reaction may include separating nucleic acid molecules hybridized with a capture probe from nucleic acid molecules not hybridized with a capture probe. The elution reaction may include separating a first subset of nucleic acid molecules from a second subset of nucleic acid molecules, the first subset of nucleic acid molecules differing from the second subset of nucleic acid molecules in average size, average GC content, genomic region, or a combination thereof.

[0119] Certain embodiments may include one or more fragmentation reactions. The fragmentation reaction may include fragmenting one or more nucleic acid molecules in a sample or subset of nucleic acid molecules to generate one or more fragmented nucleic acid molecules. The one or more nucleic acid molecules may be fragmented by sonication, needle shearing, nebulization, shearing (e.g., acoustic shearing, mechanical shearing, point-sink shearing), passage through a French pressure cell, or enzymatic digestion. Enzymatic digestion may occur by nuclease digestion (e.g., micrococcal nuclease digestion, endonuclease, exonuclease, RNAseH, or DNaseI). The fragmentation of the one or more nucleic acid molecules may result in fragment sizes of about 100 base pairs to about 2000 base pairs, about 200 base pairs to about 1500 base pairs, about 200 base pairs to about 1000 base pairs, about 200 base pairs to about 500 base pairs, about 500 base pairs to about 1500 base pairs, and about 500 base pairs to about 1000 base pairs. The one or more fragmentation reactions may result in fragments of about 50 base pairs to about 1000 base pairs in size. The one or more fragmentation reactions may result in fragment sizes of about 100 base pairs, 150 base pairs, 200 base pairs, 250 base pairs, 300 base pairs, 350 base pairs, 400 base pairs, 450 base pairs, 500 base pairs, 550 base pairs, 600 base pairs, 650 base pairs, 700 base pairs, 750 base pairs, 800 base pairs, 850 base pairs, 900 base pairs, 950 base pairs, 1000 base pairs or more.

[0120] Fragmenting the one or more nucleic acid molecules may include mechanically shearing the one or more nucleic acid molecules in the sample for a period of time, which may be for at least about 10 seconds, 15 seconds, 20 seconds, 25 seconds, 30 seconds, 35 seconds, 40 seconds, 45 seconds, 50 seconds, 55 seconds, 60 seconds, 65 seconds, 70 seconds, 75 seconds, 80 seconds, 85 seconds, 90 seconds, 95 seconds, 100 seconds, 125 seconds, 150 seconds, 175 seconds, 200 seconds, 225 seconds, 250 seconds, 275 seconds, 300 seconds, 325 seconds, 350 seconds, 375 seconds, 400 seconds, 425 seconds, 450 seconds, 475 seconds, 500 seconds, or more.

[0121] Fragmenting the one or more nucleic acid molecules can include contacting the nucleic acid sample with one or more beads.Fragmenting the one or more nucleic acid molecules can include contacting the nucleic acid sample with a plurality of beads, wherein the ratio of the volume of the plurality of beads to the volume of the nucleic acid sample is about 0.10, 0.20, 0.30, 0.40, 0.50, 0.60, 0.70, 0.80, 0.90, 1.00, 1.10, 1.20, 1.30, 1.40, 1.50, 1.60, 1.70, 1.80, 1.90, 2.00 or more. Fragmenting the one or more nucleic acid molecules can include contacting the nucleic acid sample with a plurality of beads, wherein the ratio of a volume of the plurality of beads to a volume of the nucleic acid sample is about 2.00, 1.90, 1.80, 1.70, 1.60, 1.50, 1.40, 1.30, 1.20, 1.10, 1.00, 0.90, 0.80, 0.70, 0.60, 0.50, 0.40, 0.30, 0.20, 0.10, 0.05, 0.04, 0.03, 0.02, 0.01 or less.

[0122] Certain embodiments may include performing one or more detection reactions on one or more nucleic acid molecules in a sample. The detection reaction may include one or more sequencing reactions. Alternatively, performing the detection reaction includes optical sensing, electrical sensing, or a combination thereof. The optical sensing may include optical sensing of photoluminescence photon emission, fluorescence photon emission, pyrophosphate photon emission, chemiluminescence photon emission, or a combination thereof. The electrical sensing may include electrical sensing of ion concentration, ion current modulation, nucleotide electric field, nucleotide tunneling current, or a combination thereof.

[0123] Certain embodiments may include performing one or more quantification reactions on one or more nucleic acid molecules in a sample. The quantification reactions may include sequencing, PCR, qPCR, digital PCR, or a combination thereof.

[0124] Certain embodiments may include one or more samples. Certain embodiments may include 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 20, 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75, 80, 85, 90, 95, 100 or more samples. The samples may be from a subject. The two or more samples may be from a single subject. The two or more samples may be from 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 20, 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75, 80, 85, 90, 95, 100 or more different subjects. The subject may be a mammal, a reptile, an amphibian, a bird, or a fish. A mammal may be a human, an ape, an orangutan, a monkey, a chimpanzee, a cow, a pig, a horse, a rodent, a bird, a reptile, a dog, a cat, or other animal. A reptile may be a lizard, a snake, an alligator snapping turtle, a turtle, a crocodile, a tortoise, or the like. An amphibian may be a toad, a frog, a newt, and a salamander. Examples of birds include, but are not limited to, ducks, geese, penguins, ostriches, and owls. Examples of fish include, but are not limited to, catfish, eels, sharks, and swordfish. Preferably, the subject is a human. The subject may be suffering from a disease or condition (e.g., cancer).

[0125] The two or more samples may be taken at 1 hour, 2 hours, 3 hours, 4 hours, 5 hours, 6 hours, 7 hours, 8 hours, 9 hours, 10 hours, 11 hours, 12 hours, 15 hours, 20 hours, 30 hours, 40 hours, 50 hours, 60 hours, 70 hours, 80 hours, 90 hours, 100 hours, 200 hours, 300 hours, 400 hours, 500 hours, 600 hours, 700 hours, 800 hours, 900 hours, 1000 hours or more. The time points may range from 1 hour, 2 hours, 3 hours, 4 hours, 5 hours, 6 hours, 7 hours, 8 hours, 9 hours, 10 hours, 11 hours, 12 hours, 13 hours, 14 hours, 15 hours, 16 hours, 17 hours, 18 hours, 19 hours, 20 hours, 21 hours, 22 hours, 23 hours, 24 hours, 25 hours, 30 hours, 35 hours, 40 hours, 45 hours, 50 hours, 55 hours, 60 hours or more. The time points may span 1 day, 2 days, 3 days, 4 days, 5 days, 6 days, 7 days, 8 days, 9 days, 10 days, 11 days, 12 days, 13 days, 14 days, 15 days, 16 days, 17 days, 18 days, 19 days, 20 days, 21 days, 22 days, 23 days, 24 days, 25 days, 30 days, 35 days, 40 days, 45 days, 50 days, 55 days, 60 days or more. The time points may range from 1 week, 2 weeks, 3 weeks, 4 weeks, 5 weeks, 6 weeks, 7 weeks, 8 weeks, 9 weeks, 10 weeks, 11 weeks, 12 weeks, 13 weeks, 14 weeks, 15 weeks, 16 weeks, 17 weeks, 18 weeks, 19 weeks, 20 weeks, 21 weeks, 22 weeks, 23 weeks, 24 weeks, 25 weeks, 30 weeks, 35 weeks, 40 weeks, 45 weeks, 50 weeks, 55 weeks, 60 weeks or more. The time points may span 1 month, 2 months, 3 months, 4 months, 5 months, 6 months, 7 months, 8 months, 9 months, 10 months, 11 months, 12 months, 13 months, 14 months, 15 months, 16 months, 17 months, 18 months, 19 months, 20 months, 21 months, 22 months, 23 months, 24 months, 25 months, 30 months, 35 months, 40 months, 45 months, 50 months, 55 months, 60 months or more.The time point may span 1 year, 2 years, 3 years, 4 years, 5 years, 6 years, 7 years, 8 years, 9 years, 10 years, 11 years, 12 years, 13 years, 14 years, 15 years, 16 years, 17 years, 18 years, 19 years, 20 years, 21 years, 22 years, 23 years, 24 years, 25 years, 30 years, 35 years, 40 years, 45 years, 50 years, 55 years, 60 years or more.

[0126] The sample may be from a body fluid, cell, skin, tissue, organ, or combination thereof. The sample may be blood, plasma, blood fraction, saliva, sputum, urine, semen, vaginal fluid, cerebrospinal fluid, feces, cell, or tissue biopsy. The sample may be from the adrenal gland, adnexa, bladder, brain, ear, esophagus, eye, gallbladder, heart, kidney, large intestine, liver, lung, oral cavity, muscle, nasal cavity, pancreas, parathyroid gland, pineal gland, pituitary gland, skin, small intestine, spleen, stomach, thymus, thyroid gland, trachea, uterus, appendix, cornea, skin, heart valve, artery, or vein.

[0127] The sample may include one or more nucleic acid molecules. The nucleic acid molecules may be DNA molecules, RNA molecules (e.g., mRNA, cRNA, or miRNA), and DNA / RNA hybrids. Examples of DNA molecules include, but are not limited to, double-stranded DNA, single-stranded DNA, single-stranded DNA hairpins, cDNA, and genomic DNA. The nucleic acids may be RNA molecules, such as double-stranded RNA, single-stranded RNA, ncRNA, RNA hairpins, and mRNA. Examples of ncRNA include, but are not limited to, siRNA, miRNA, snoRNA, piRNA, tiRNA, PASR, TASR, aTASR, TSSa-RNA, snRNA, RE-RNA, uaRNA, x-ncRNA, hY RNA, usRNA, snaR, and vtRNA.

[0128] Certain embodiments may include one or more containers. Certain embodiments may include 1 or more, 2 or more, 3 or more, 4 or more, 5 or more, 6 or more, 7 or more, 8 or more, 9 or more, 10 or more, 20 or more, 30 or more, 40 or more, 50 or more, 60 or more, 70 or more, 80 or more, 90 or more, 100 or more, 125 or more, 150 or more, 175 or more, 200 or more, 250 or more, 300 or more, 350 or more, 400 or more, 500 or more, 600 or more, 700 or more, 800 or more, 900 or more, or 1000 or more containers. The one or more containers may be different, similar, identical, or combinations thereof. Examples of containers include, but are not limited to, plates, microplates, PCR plates, wells, microwells, tubes, eppendorf tubes, vials, arrays, microarrays, and chips.

[0129] Certain embodiments may include one or more reagents. Certain embodiments may include 1 or more, 2 or more, 3 or more, 4 or more, 5 or more, 6 or more, 7 or more, 8 or more, 9 or more, 10 or more, 20 or more, 30 or more, 40 or more, 50 or more, 60 or more, 70 or more, 80 or more, 90 or more, 100 or more, 125 or more, 150 or more, 175 or more, 200 or more, 250 or more, 300 or more, 350 or more, 400 or more, 500 or more, 600 or more, 700 or more, 800 or more, 900 or more, or 1000 or more reagents. The one or more reagents may be different, similar, identical, or a combination thereof. The reagent may improve the efficiency of one or more assays. The reagent may improve the stability of the nucleic acid molecule or variant or derivative thereof. Reagents may include, but are not limited to, enzymes, proteases, nucleases, molecules, polymerases, reverse transcriptases, ligases, and chemical compounds. Certain embodiments may include performing an assay that includes one or more antioxidants. Generally, an antioxidant is a molecule that inhibits the oxidation of another molecule. Examples of antioxidants include, but are not limited to, ascorbic acid (e.g., vitamin C), glutathione, lipoic acid, uric acid, carotene, alpha-tocopherol (e.g., vitamin E), ubiquinol (e.g., coenzyme Q), and vitamin A.

[0130] Certain embodiments may include one or more buffers or solutions. The one or more buffers or solutions may be different, similar, identical, or a combination thereof. The buffers or solutions may improve the efficiency of one or more assays. The buffers or solutions may improve the stability of the nucleic acid molecule or variants or derivatives thereof. The buffers or solutions may include, but are not limited to, wash buffers, elution buffers, and hybridization buffers.

[0131] Certain embodiments may include one or more beads, multiple beads, or one or more bead sets. Certain embodiments may include one or more, two or more, three or more, four or more, five or more, six or more, seven or more, eight or more, nine or more, ten or more, twenty or more, thirty or more, forty or more, fifty or more, sixty or more, seventy or more, eighty or more, ninety or more, one hundred or more, one or more beads or bead sets. The one or more beads or bead sets may be different, similar, identical, or combinations thereof. The beads may be magnetic, antibody coated, protein A cross-linked, protein G cross-linked, streptavidin coated, oligonucleotide conjugated, silica coated, or combinations thereof. Examples of beads include, but are not limited to, AMPure® beads, AMPure® XP beads, streptavidin beads, agarose beads, magnetic beads, Dynabeads®, MACS® microbeads, antibody-bound beads (e.g., anti-immunoglobulin microbeads), protein A-bound beads, protein G-bound beads, protein A / G-bound beads, protein L-bound beads, oligo dT-bound beads, silica beads, silica-like beads, anti-biotin microbeads, anti-fluorescent chromium microbeads, and BcMag™ carboxy-terminated magnetic beads. In some embodiments of the present disclosure, the one or more beads comprise one or more AMPure® beads. Alternatively or additionally, the one or more beads comprise AMPure® XP beads.

[0132] Certain embodiments may include one or more primers, multiple primers, or one or more primer sets. The primers may further include one or more linkers. The primers may further include one or more labels. The primers may be used in one or more assays. For example, the primers are used in one or more sequencing reactions, amplification reactions, or combinations thereof. Certain embodiments may include one or more primers or primer sets, including 1 or more, 2 or more, 3 or more, 4 or more, 5 or more, 6 or more, 7 or more, 8 or more, 9 or more, 10 or more, 20 or more, 30 or more, 40 or more, 50 or more, 60 or more, 70 or more, 80 or more, 90 or more, 100 or more, 125 or more, 150 or more, 175 or more, 200 or more, 250 or more, 300 or more, 350 or more, 400 or more, 500 or more, 600 or more, 700 or more, 800 or more, 900 or more, 1000 or more. The primers may include about 100 nucleotides. The primer may comprise about 10 to about 500 nucleotides, about 20 to about 450 nucleotides, about 30 to about 400 nucleotides, about 40 to about 350 nucleotides, about 50 to about 300 nucleotides, about 60 to about 250 nucleotides, about 70 to about 200 nucleotides, or about 80 to about 150 nucleotides. In some embodiments of the present disclosure, the primer comprises about 80 nucleotides to about 100 nucleotides. The one or more primers or primer sets may be different, similar, identical, or a combination thereof.

[0133] The primers hybridize to at least a portion of one or more nucleic acid molecules or variants or derivatives thereof in a sample or subset of nucleic acid molecules. The primers may hybridize to one or more genomic regions. The primers may hybridize to different, similar and / or identical genomic regions. The one or more primers may be at least about 50%, 55%, 60%, 65%, 70%, 75%, 80%, 85%, 90%, 95%, 97%, 99% or more complementary to the one or more nucleic acid molecules or variants or derivatives thereof.

[0134] The primer may comprise one or more nucleotides. The primer may comprise 1 or more, 2 or more, 3 or more, 4 or more, 5 or more, 6 or more, 7 or more, 8 or more, 9 or more, 10 or more, 20 or more, 30 or more, 40 or more, 50 or more, 60 or more, 70 or more, 80 or more, 90 or more, 100 or more, 125 or more, 150 or more, 175 or more, 200 or more, 250 or more, 300 or more, 350 or more, 400 or more, 500 or more, 600 or more, 700 or more, 800 or more, 900 or more, or 1000 or more nucleotides. The primer may comprise about 100 nucleotides. The primer may comprise about 10 to about 500 nucleotides, about 20 to about 450 nucleotides, about 30 to about 400 nucleotides, about 40 to about 350 nucleotides, about 50 to about 300 nucleotides, about 60 to about 250 nucleotides, about 70 to about 200 nucleotides, or about 80 to about 150 nucleotides. In some embodiments of the present disclosure, the primer comprises about 80 nucleotides to about 100 nucleotides.

[0135] The plurality or primer set may include two or more primers having the same, similar, and / or different sequences, linkers, and / or labels. For example, the two or more primers include the same sequence. In another example, the two or more primers include similar sequences. In yet another example, the two or more primers include different sequences. The two or more primers may further include one or more linkers. The two or more primers may further include different linkers. The two or more primers may further include similar linkers. The two or more primers may further include identical linkers. The two or more primers may further include one or more labels. The two or more primers may further include different labels. The two or more primers may further include similar labels. The two or more primers may further include the same labels.

[0136] The capture probes, primers, labels, and / or beads may comprise one or more nucleotides that may include RNA, DNA, a mixture of DNA and RNA residues, or modified analogs such as 2'-0Me, 2'-fluoro (2'-F), locked nucleic acid (LNA), or abasic sites.

[0137] Certain embodiments may include one or more labels. Certain embodiments may include one or more, 1 or more, 2 or more, 3 or more, 4 or more, 5 or more, 6 or more, 7 or more, 8 or more, 9 or more, 10 or more, 20 or more, 30 or more, 40 or more, 50 or more, 60 or more, 70 or more, 80 or more, 90 or more, 100 or more, 125 or more, 150 or more, 175 or more, 200 or more, 250 or more, 300 or more, 350 or more, 400 or more, 500 or more, 600 or more, 700 or more, 800 or more, 900 or more, 1000 or more labels. The one or more labels may be different, similar, identical, or a combination thereof.

[0138] Examples of labels include, but are not limited to, chemical, biochemical, biological, colorimetric, enzymatic, fluorescent, and luminescent labels known in the art, including dyes, photocrosslinkers, cytotoxic compounds, drugs, affinity labels, photoaffinity labels, reactive compounds, antibodies or antibody fragments, biomaterials, nanoparticles, spin labels, fluorophores, metal-containing moieties, radioactive moieties, novel functional groups, groups that interact covalently or noncovalently with other molecules, photocaged moieties, actinically excitable moieties, ligands, photoisomerizable moieties, biotin, biotin analogs, moieties incorporating heavy atoms, chemically cleavable groups, photocleavable groups, redox-active agents, isotopically labeled moieties, biophysical probes, phosphorescent groups, chemiluminescent groups, electron-dense groups, magnetic groups, intercalating groups, chromophores, energy transfer agents, biologically active agents, detectable labels, or combinations thereof.

[0139] The label may be a chemical label. Examples of chemical labels may include, but are not limited to, biotin and radioisotopes (e.g., iodine, carbon, phosphate, hydrogen).

[0140] The methods, kits, and compositions disclosed herein may include biological labels, including, but not limited to, metabolic labels, including bioorthogonal azide-modified amino acids, sugars, and other compounds.

[0141] The methods, kits, and compositions disclosed herein may include an enzyme label, including, but not limited to, horseradish peroxidase (HRP), alkaline phosphatase (AP), glucose oxidase, and O-galactosidase. The enzyme label may be luciferase.

[0142] The methods, kits, and compositions disclosed herein may include a biological label. The fluorescent label may be an organic dye (e.g., FITC), a biological fluorophore (e.g., green fluorescent protein), or a quantum dot. A non-limiting list of fluorescent labels includes fluorescein isothiocyanate (FITC), DyLight® Fluor, fluorescein, rhodamine (tetramethylrhodamine isothiocyanate, TRITC), coumarin, lucifer yellow, and BODIPY. The label may be a fluorophore. Examples of fluorophores include indocarbocyanine (C3), indodicarbocyanine (C5), Cy3, Cy3.5, Cy5, Cy5.5, Cy7, Texas Red, Pacific Blue, Oregon Green 488, Alexa Fluor® 355, Alexa Fluor® 488, Alexa Fluor® 532, Alexa Fluor® 546, Alexa Fluor® 555, Alexa Fluor® 568, Alexa Fluor® 594, Alexa Fluor® 647, Alexa Fluor® 660, Alexa Fluor® 570, Alexa Fluor® 572, Alexa Fluor® 576, Alexa Fluor® 578, Alexa Fluor® 579 ... Fluorescent labels include, but are not limited to, Fluor® 680, JOE, Lissamine, rhodamine green, BODIPY, fluorescein isothiocyanate (FITC), carboxyfluorescein (FAM), phycoerythrin, rhodamine, dichlororhodamine (dRhodamine), carboxytetramethylrhodamine (TAMRA), carboxy-X-rhodamine (ROX™), LIZ™, VIC™, NED™, PET™, SYBR, PicoGreen®, RiboGreen®, etc. The fluorescent label can be green fluorescent protein (GFP), red fluorescent protein (RFP), yellow fluorescent protein, phycobiliproteins (e.g., allophycocyanin, phycocyanin, phycoerythrin, and phycoerythrocyanin).

[0143] Certain embodiments may include one or more linkers. Certain embodiments may include one or more, 2 or more, 3 or more, 4 or more, 5 or more, 6 or more, 7 or more, 8 or more, 9 or more, 10 or more, 20 or more, 30 or more, 40 or more, 50 or more, 60 or more, 70 or more, 80 or more, 90 or more, 100 or more, 125 or more, 150 or more, 175 or more, 200 or more, 250 or more, 300 or more, 350 or more, 400 or more, 500 or more, 600 or more, 700 or more, 800 or more, 900 or more, 1000 or more linkers. The one or more linkers may be different, similar, identical, or a combination thereof.

[0144] Suitable linkers include any chemical or biological compound that can be attached to a label, primer, and / or capture probe as disclosed herein. When the linker is attached to both a label and a primer or capture probe, a suitable linker allows sufficient separation of the label and the primer or capture probe. A suitable linker does not significantly interfere with the ability of the primer and / or capture probe to hybridize to a nucleic acid molecule, a portion thereof, or a variant or derivative thereof. A suitable linker does not significantly interfere with the ability of the label to be detected. The linker can be rigid. The linker can be flexible. The linker can be semi-rigid. The linker can be proteolytically stable (e.g., resistant to proteolytic cleavage). The linker can be proteolytically unstable (e.g., susceptible to proteolytic cleavage). The linker can be helical. The linker can be non-helical. The linker can be coiled. The linker can be triple-stranded. The linker can include a turn conformation. The linker may be a single chain. The linker may be a long chain. The linker may be a short chain. The linker may comprise at least about 5 residues, at least about 10 residues, at least about 15 residues, at least about 20 residues, at least about 25 residues, at least about 30 residues, or at least about 40 residues or more.

[0145] Examples of linkers include, but are not limited to, hydrazones, disulfides, thioethers, and peptide linkers. The linker can be a peptide linker. The peptide linker can include a proline residue. The peptide linker can include arginine, phenylalanine, threonine, glutamine, glutamic acid, or any combination thereof. The linker can be a heterobifunctional crosslinker.

[0146] Certain embodiments include performing 1 or more, 2 or more, 3 or more, 4 or more, 5 or more, 6 or more, 7 or more, 8 or more, 9 or more, 10 or more, 11 or more, 12 or more, 13 or more, 14 or more, 15 or more, 20 or more, 25 or more, 30 or more, 35 or more, 40 or more, 45 or more, or 50 or more assays on a sample that includes one or more nucleic acid molecules. The two or more assays can be different, similar, identical, or a combination thereof. For example, certain embodiments include performing two or more sequencing reactions. In another example, certain embodiments include performing two or more assays, where at least one of the two or more assays includes a sequencing reaction. In yet another example, certain embodiments include performing two or more assays, where at least two of the two or more assays include a sequencing reaction and a hybridization reaction. The two or more assays may be performed sequentially, simultaneously, or a combination thereof. For example, the two or more sequencing reactions may be performed simultaneously. In another example, certain embodiments include performing a hybridization reaction followed by a sequencing reaction. In yet another example, certain embodiments include performing two or more hybridization reactions simultaneously, followed by performing two or more sequencing reactions simultaneously. The two or more assays may be performed by one or more devices. For example, the two or more amplification reactions may be performed by a PCR machine. In another example, the two or more sequencing reactions may be performed by two or more sequencers.

[0147] C. Device Certain embodiments may include one or more devices. Certain embodiments may include one or more assays that include one or more devices. Certain embodiments may include the use of one or more devices to perform one or more steps or assays. Certain embodiments may include the use of one or more devices in one or more steps or assays. For example, performing a sequencing reaction may include one or more sequencers. In another example, generating a subset of nucleic acid molecules may include the use of one or more magnetic separators. In yet another example, one or more processors may be used in the analysis of one or more nucleic acid samples. Examples of devices include, but are not limited to, sequencers, thermocyclers, real-time PCR instruments, magnetic separators, transmission devices, hybridization chambers, electrophoresis devices, centrifuges, microscopes, imagers, fluorometers, luminometers, plate readers, computers, processors, and bioanalyzers.

[0148] Certain embodiments may include one or more sequencers. The one or more sequencers may include one or more HiSeq, MiSeq, HiScan, Genome Analyzer IIx, SOLiD Sequencer, Ion Torrent PGM, 454 GS Junior, Pac Bio RS, or combinations thereof. The one or more sequencers may include one or more sequencing platforms. The one or more sequencing platforms may include GS FLX 454 by Life Technologies / Roche, Genome Analyzer by Solexa / Illumina, SOLiD by Applied Biosystems, CGA Platform by Complete Genomics, PacBio RS by Pacific Biosciences, or combinations thereof.

[0149] Certain embodiments may include one or more thermocyclers. The one or more thermocyclers may be used for amplifying one or more nucleic acid molecules. Certain embodiments may include one or more real-time PCR instruments. The one or more real-time PCR instruments may include a thermal cycler and a fluorometer. The one or more thermocyclers may be used for amplifying and detecting one or more nucleic acid molecules.

[0150] Certain embodiments may include one or more magnetic separators, which may be used to separate paramagnetic and ferromagnetic particles from the suspension. The one or more magnetic separators may include one or more LifeStep™ biomagnetic separators, SPHERO™ FlexiMag separators, SPHERO™ MicroMag separators, SPHERO™ HandiMag separators, SPHERO™ MiniTube Mag separators, SPHERO™ UltraMag separators, DynaMag™ magnets, DynaMag™-2 magnets, or combinations thereof.

[0151] Certain embodiments may include one or more bioanalyzers. Generally, a bioanalyzer is a chip-based capillary electrophoresis device capable of analyzing RNA, DNA, and proteins. The one or more bioanalyzers may include an Agilent 2100 bioanalyzer.

[0152] Certain embodiments may include one or more processors. The one or more processors may analyze, compile, store, sort, combine, evaluate, or otherwise process one or more data and / or results from one or more assays, one or more data and / or results based on or derived from one or more assays, one or more outputs from one or more assays, one or more outputs based on or derived from one or more data and / or results, one or more outputs based on or derived from one or more data and / or results, or combinations thereof. The one or more processors may transmit one or more data, results, or outputs from one or more assays, one or more data, results, or outputs based on or derived from one or more assays, one or more outputs from one or more data or results, one or more outputs based on or derived from one or more data or results, or combinations thereof. The one or more processors may receive and / or store requests from a user. The one or more processors may create or generate one or more data, results, outputs. The one or more processors may generate or generate one or more biomedical reports. The one or more processors may transmit one or more biomedical reports. The one or more processors may analyze, compile, store, sort, combine, evaluate, or otherwise process information from one or more databases, one or more data or results, one or more outputs, or a combination thereof. The one or more processors may analyze, compile, store, sort, combine, evaluate, or otherwise process information from 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 30 or more databases. The one or more processors may transmit one or more requests, data, results, outputs, and / or information to one or more users, processors, computers, computer systems, memory locations, devices, databases, or combinations thereof.The one or more processors may receive one or more requests, data, results, output, and / or information from one or more users, processors, computers, computer systems, memory locations, devices, databases, or combinations thereof. The one or more processors may retrieve one or more requests, data, results, output, and / or information from one or more users, processors, computers, computer systems, memory locations, devices, databases, or combinations thereof.

[0153] Certain embodiments may include one or more memory locations that may store information, data, results, output, requests, or combinations thereof. The one or more memory locations may receive information, data, results, output, requests, or combinations thereof from one or more users, processors, computers, computer systems, devices, or combinations thereof.

[0154] The methods described herein may be implemented using one or more computers and / or computer systems. A computer or computer system may include electronic storage locations (e.g., databases, memory) with machine-executable code for implementing the methods provided herein, and one or more processors for executing the machine-executable code.

[0155] The code may be precompiled and configured for use on a device with a processor adapted to execute the code, or it may be compiled at run time. The code may be provided in a selectable programming language such that the code can be executed in a precompiled or compiled form.

[0156] The one or more computers and / or computer systems may analyze, compile, store, sort, combine, evaluate, or otherwise process one or more data and / or results from one or more assays, one or more data and / or results based on or derived from one or more assays, one or more outputs from one or more assays, one or more outputs based on or derived from one or more data and / or results, one or more outputs based on or derived from one or more data and / or results, or a combination thereof. The one or more computers and / or computer systems may transmit one or more data, results, or outputs from one or more assays, one or more data, results, or outputs based on or derived from one or more assays, one or more outputs from one or more data or results, one or more outputs based on or derived from one or more data or results, or a combination thereof. The one or more computers and / or computer systems may receive and / or store requests from a user. The one or more computers and / or computer systems may create or generate one or more data, results, outputs. The one or more computers and / or computer systems may generate or generate one or more biomedical reports. The one or more computers and / or computer systems may transmit one or more biomedical reports. The one or more computers and / or computer systems may analyze, compile, store, sort, combine, evaluate, or otherwise process information from one or more databases, one or more data or results, one or more outputs, or a combination thereof. The one or more computers and / or computer systems may analyze, compile, store, sort, combine, evaluate, or otherwise process information from 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 30 or more databases.The one or more computers and / or computer systems may transmit one or more requests, data, results, output and / or information to one or more users, processors, computers, computer systems, memory locations, devices, or combinations thereof. The one or more computers and / or computer systems may receive one or more requests, data, results, output and / or information from one or more users, processors, computers, computer systems, memory locations, devices, or combinations thereof. The one or more computers and / or computer systems may retrieve one or more requests, data, results, output and / or information from one or more users, processors, computers, computer systems, memory locations, devices, databases, or combinations thereof.

[0157] D. Database Certain embodiments may include one or more databases. Certain embodiments may include at least about 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 30 or more databases. The databases may include genomic databases, proteomic databases, pharmacogenomic databases, biomedical databases, and scientific databases. The databases may be public databases. Alternatively or additionally, the databases may include proprietary databases. The databases may be commercial databases. Such databases include, but are not limited to, Cosmic, GnomAD, Dbsnp, Mills Indels, MendelDB, PharmGKB, Varimed, Regulome, BreakSeq (curated junctions), OMIM (Online Mendelian Inheritance in Man), HGMD (Human Genome Mutation Database), NCBI db SNP, NCBI RefSeq, GENCODE, GO (Gene Ontology), and KEGG (Kyoto Encyclopedia of Genes and Genomes).

[0158] Certain embodiments may include analyzing one or more databases. Certain embodiments may include analyzing at least about 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 30 or more databases. Analyzing the one or more databases may include one or more algorithms, computers, processors, memory locations, devices, or combinations thereof.

[0159] Certain embodiments may include identifying one or more nucleic acid regions based on data and / or information from one or more databases. Certain embodiments may include identifying a set of one or more nucleic acid regions based on data and / or information from one or more databases. Certain embodiments may include identifying one or more nucleic acid regions and / or sets of nucleic acid regions based on data and / or information from at least about two or more databases. Certain embodiments may include identifying one or more nucleic acid regions and / or sets of nucleic acid regions based on data and / or information from at least about three or more databases. Certain embodiments may include identifying one or more nucleic acid regions and / or sets of nucleic acid regions based on data and / or information from at least about 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 30 or more databases.

[0160] Certain embodiments may include analyzing one or more results based on data and / or information from one or more databases. Certain embodiments may include analyzing one or more sets of results based on data and / or information from one or more databases. Certain embodiments may include analyzing one or more combined results based on data and / or information from one or more databases. Certain embodiments may include analyzing one or more results, sets of results, combined results based on data and / or information from at least about two or more databases. Certain embodiments may include analyzing one or more results, sets of results, combined results based on data and / or information from at least about three or more databases. Certain embodiments may include analyzing one or more results, sets of results, combined results based on data and / or information from at least about 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 30 or more databases.

[0161] Certain embodiments may include comparing one or more results based on data and / or information from one or more databases. Certain embodiments may include comparing one or more sets of results based on data and / or information from one or more databases. Certain embodiments may include comparing one or more combined results based on data and / or information from one or more databases. Certain embodiments may include comparing one or more results, sets of results, combined results based on data and / or information from at least about two or more databases. Certain embodiments may include comparing one or more results, sets of results, combined results based on data and / or information from at least about three or more databases. Certain embodiments may include comparing one or more results, sets of results, combined results based on data and / or information from at least about 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 30 or more databases.

[0162] Certain embodiments may include biomedical databases, genomic databases, biomedical reports, disease reports, case-control analyses, and rare variant detection analyses based on data and / or information from one or more databases, one or more assays, one or more data or results, one or more outputs based on or derived from one or more assays, one or more outputs based on or derived from one or more data or results, or combinations thereof.

[0163] E. Datasets and Analysis Certain embodiments may include one or more data, one or more datasets, one or more combined data, one or more combined datasets, one or more results, one or more sets of results, one or more combined results, or a combination thereof. The data and / or results may be based on or derived from one or more assays, one or more databases, or a combination thereof. Certain embodiments may include analysis of one or more data, one or more datasets, one or more combined data, one or more combined datasets, one or more results, one or more sets of results, one or more combined results, or a combination thereof. Certain embodiments may include processing of one or more data, one or more datasets, one or more combined data, one or more combined datasets, one or more results, one or more sets of results, one or more combined results, or a combination thereof.

[0164] Certain embodiments may include at least one analysis and at least one processing of one or more data, one or more datasets, one or more combined data, one or more combined datasets, one or more results, one or more sets of results, one or more combined results, or a combination thereof. Certain embodiments may include one or more analyses and one or more processing of one or more data, one or more datasets, one or more combined data, one or more combined datasets, one or more results, one or more sets of results, one or more combined results, or a combination thereof. Certain embodiments may include at least 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 15, 20, 30, 40, 50, 60, 70, 80, 90, 100, 200, 300, 400, 500, 600, 700, 800, 900, 1000 or more different analyses of one or more data, one or more datasets, one or more combined data, one or more combined datasets, one or more results, one or more sets of results, one or more combined results, or combinations thereof. Certain embodiments may include at least 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 15, 20, 30, 40, 50, 60, 70, 80, 90, 100, 200, 300, 400, 500, 600, 700, 800, 900, 1000 or more different processing of one or more data, one or more data sets, one or more combined data, one or more combined data sets, one or more results, one or more sets of results, one or more combined results, or combinations thereof. The one or more analyses and / or one or more processing may be performed simultaneously, sequentially, or a combination thereof.

[0165] The one or more analyses and / or one or more treatments may be performed at 1 hour, 2 hours, 3 hours, 4 hours, 5 hours, 6 hours, 7 hours, 8 hours, 9 hours, 10 hours, 11 hours, 12 hours, 15 hours, 20 hours, 30 hours, 40 hours, 50 hours, 60 hours, 70 hours, 80 hours, 90 hours, 100 hours, 200 hours, 300 hours, 400 hours, 500 hours, 600 hours, 700 hours, 800 hours, 900 hours, 1000 hours or more. The time points may range from 1 hour, 2 hours, 3 hours, 4 hours, 5 hours, 6 hours, 7 hours, 8 hours, 9 hours, 10 hours, 11 hours, 12 hours, 13 hours, 14 hours, 15 hours, 16 hours, 17 hours, 18 hours, 19 hours, 20 hours, 21 hours, 22 hours, 23 hours, 24 hours, 25 hours, 30 hours, 35 hours, 40 hours, 45 hours, 50 hours, 55 hours, 60 hours or more. The time points may span 1 day, 2 days, 3 days, 4 days, 5 days, 6 days, 7 days, 8 days, 9 days, 10 days, 11 days, 12 days, 13 days, 14 days, 15 days, 16 days, 17 days, 18 days, 19 days, 20 days, 21 days, 22 days, 23 days, 24 days, 25 days, 30 days, 35 days, 40 days, 45 days, 50 days, 55 days, 60 days or more. The time points may range from 1 week, 2 weeks, 3 weeks, 4 weeks, 5 weeks, 6 weeks, 7 weeks, 8 weeks, 9 weeks, 10 weeks, 11 weeks, 12 weeks, 13 weeks, 14 weeks, 15 weeks, 16 weeks, 17 weeks, 18 weeks, 19 weeks, 20 weeks, 21 weeks, 22 weeks, 23 weeks, 24 weeks, 25 weeks, 30 weeks, 35 weeks, 40 weeks, 45 weeks, 50 weeks, 55 weeks, 60 weeks or more. The time points may span 1 month, 2 months, 3 months, 4 months, 5 months, 6 months, 7 months, 8 months, 9 months, 10 months, 11 months, 12 months, 13 months, 14 months, 15 months, 16 months, 17 months, 18 months, 19 months, 20 months, 21 months, 22 months, 23 months, 24 months, 25 months, 30 months, 35 months, 40 months, 45 months, 50 months, 55 months, 60 months or more.The time point may span 1 year, 2 years, 3 years, 4 years, 5 years, 6 years, 7 years, 8 years, 9 years, 10 years, 11 years, 12 years, 13 years, 14 years, 15 years, 16 years, 17 years, 18 years, 19 years, 20 years, 21 years, 22 years, 23 years, 24 years, 25 years, 30 years, 35 years, 40 years, 45 years, 50 years, 55 years, 60 years or more.

[0166] Certain embodiments may include one or more data. The one or more data may include one or more raw data based on or derived from one or more assays. The one or more data may include one or more raw data based on or derived from one or more databases. The one or more data may include at least partially analyzed data based on or derived from one or more raw data. The one or more data may include at least partially processed data based on or derived from one or more raw data. The one or more data may include fully analyzed data based on or derived from one or more raw data. The one or more data may include fully processed data based on or derived from one or more raw data. The data may include sequencing read data or expression data. The data may include biomedical, chemical, pharmacological, and / or genetic information.

[0167] Certain embodiments may include one or more combined data. The one or more combined data may include two or more data. The one or more combined data may include two or more data sets. The one or more combined data may include one or more raw data based on or derived from one or more assays. The one or more combined data may include one or more raw data based on or derived from one or more databases. The one or more combined data may include at least partially analyzed data based on or derived from one or more raw data. The one or more combined data may include at least partially processed data based on or derived from one or more raw data. The one or more combined data may include fully analyzed data based on or derived from one or more raw data. The one or more combined data may include fully processed data based on or derived from one or more raw data. The one or more combined data may include sequencing read data or expression data. The one or more combined data may include biomedical, chemical, pharmacological, and / or genetic information.

[0168] Certain embodiments may include one or more datasets. The one or more datasets may include one or more data. The one or more datasets may include one or more combined data. The one or more datasets may include one or more raw data based on or derived from one or more assays. The one or more datasets may include one or more raw data based on or derived from one or more databases. The one or more datasets may include at least partially analyzed data based on or derived from one or more raw data. The one or more datasets may include at least partially processed data based on or derived from one or more raw data. The one or more datasets may include fully analyzed data based on or derived from one or more raw data. The one or more datasets may include fully processed data based on or derived from one or more raw data. The datasets may include sequencing read data or expression data. The datasets may include biomedical, scientific, pharmacological, and / or genetic information.

[0169] Certain embodiments may include one or more combined datasets. The one or more combined datasets may include two or more data. The one or more combined datasets may include two or more combined data. The one or more combined datasets may include two or more datasets. The one or more combined datasets may include one or more raw data based on or derived from one or more assays. The one or more combined datasets may include one or more raw data based on or derived from one or more databases. The one or more combined datasets may include at least partially analyzed data based on or derived from one or more raw data. The one or more combined datasets may include at least partially processed data based on or derived from one or more raw data. The one or more combined datasets may include fully analyzed data based on or derived from one or more raw data. The one or more combined datasets may include fully processed data based on or derived from one or more raw data. Certain embodiments may further include further processing and / or analysis of the combined datasets. The one or more combined datasets may include sequencing read data or expression data. The one or more combined datasets may include biomedical, chemical, pharmacological, and / or genetic information.

[0170] Certain embodiments may include one or more results. The one or more results may include one or more data, datasets, combined data, and / or combined datasets. The one or more results may be based on or derived from one or more data, datasets, combined data, and / or combined datasets. The one or more results may be generated from one or more assays. The one or more results may be based on or derived from one or more assays. The one or more results may be based on or derived from one or more databases. The one or more results may include at least partially analyzed results based on or derived from one or more data, datasets, combined data, and / or combined datasets. The one or more results may include at least partially processed results based on or derived from one or more data, datasets, combined data, and / or combined datasets. The one or more results may include fully analyzed results based on or derived from one or more data, datasets, combined data, and / or combined datasets. The one or more results may include one or more data, datasets, combined data, and / or fully processed results based on or derived from combined datasets. The results may include sequencing read data or expression data. The results may include biomedical, chemical, pharmacological, and / or genetic information.

[0171] Certain embodiments may include one or more result sets. The one or more result sets may include one or more data, datasets, combined data, and / or combined datasets. The one or more result sets may be based on or derived from one or more data, datasets, combined data, and / or combined datasets. The one or more result sets may be generated from one or more assays. The one or more result sets may be based on or derived from one or more assays. The one or more result sets may be based on or derived from one or more databases. The one or more result sets may include an at least partially analyzed result set based on or derived from one or more data, datasets, combined data, and / or combined datasets. The one or more result sets may include an at least partially processed result set based on or derived from one or more data, datasets, combined data, and / or combined datasets. The one or more result sets may include a fully analyzed result set based on or derived from one or more data, datasets, combined data, and / or combined datasets. The one or more result sets may include one or more data, datasets, combined data, and / or fully processed result sets based on or derived from combined datasets. The result sets may include sequencing read data or expression data. The result sets may include biomedical, chemical, pharmacological, and / or genetic information.

[0172] Certain embodiments may include one or more combined results. The combined results may include one or more results, sets of results, and / or sets of combined results. The combined results may be based on or derived from one or more results, sets of results, and / or sets of combined results. The one or more combined results may include one or more data, datasets, combined data, and / or combined datasets. The one or more combined results may be based on or derived from one or more data, datasets, combined data, and / or combined datasets. The one or more combined results may be generated from one or more assays. The one or more combined results may be based on or derived from one or more assays. The one or more combined results may be based on or derived from one or more databases. The one or more combined results may include at least partially analyzed combined results based on or derived from one or more data, datasets, combined data, and / or combined datasets. The one or more combined results may include at least partially processed combined results based on or derived from one or more data, datasets, combined data, and / or combined datasets. The one or more combined results may include fully analyzed combined results based on or derived from one or more data, datasets, combined data, and / or combined datasets. The one or more combined results may include fully processed combined results based on or derived from one or more data, datasets, combined data, and / or combined datasets. The combined results may include sequencing read data or expression data. The combined results may include biomedical, chemical, pharmacological, and / or genetic information.

[0173] Certain embodiments may include one or more combined result sets. The combined result set may include one or more results, result sets, and / or combined results. The combined result set may be based on or derived from one or more results, result sets, and / or combined results. The one or more combined result sets may include one or more data, data sets, combined data, and / or combined datasets. The one or more combined result sets may be based on or derived from one or more data, data sets, combined data, and / or combined datasets. The one or more combined result sets may be generated from one or more assays. The one or more combined result sets may be based on or derived from one or more assays. The one or more combined result sets may be based on or derived from one or more databases. The one or more combined result sets may include an at least partially analyzed combined result set based on or derived from one or more data, data sets, combined data, and / or combined datasets. The one or more combined result sets may include at least partially processed combined result sets based on or derived from one or more data, datasets, combined data, and / or combined datasets. The one or more combined result sets may include fully analyzed combined result sets based on or derived from one or more data, datasets, combined data, and / or combined datasets. The one or more combined result sets may include fully processed combined result sets based on or derived from one or more data, datasets, combined data, and / or combined datasets. The combined result sets may include sequencing read data or expression data. The combined result sets may include biomedical, chemical, pharmacological, and / or genetic information.

[0174] Certain embodiments may include one or more outputs, sets of outputs, combined outputs, and / or sets of combined outputs. The methods, libraries, kits, and systems described herein may include generating one or more outputs, sets of outputs, combined outputs, and / or sets of combined outputs. The set of outputs may include one or more outputs, one or more combined outputs, or a combination thereof. The combined output may include one or more outputs, one or more sets of outputs, one or more sets of combined outputs, or a combination thereof. The combined output may include one or more outputs, one or more sets of outputs, one or more combined outputs, or a combination thereof. The one or more outputs, sets of outputs, combined outputs, and / or sets of combined outputs may be based on or derived from one or more data, one or more datasets, one or more combined data, one or more combined datasets, one or more results, one or more sets of results, one or more combined results, or a combination thereof. The one or more outputs, sets of outputs, combined outputs, and / or sets of combined outputs may be based on or derived from one or more databases. The one or more outputs, sets of outputs, combined outputs, and / or sets of combined outputs may include one or more biomedical reports, biomedical outputs, rare variant outputs, pharmacogenetic outputs, population study outputs, case-control outputs, biomedical databases, genomic databases, disease databases, net content.

[0175] Certain embodiments may include one or more biomedical outputs, one or more sets of biomedical outputs, one or more combined biomedical outputs, one or more sets of combined biomedical outputs. The methods, libraries, kits, and systems described herein may include generating one or more biomedical outputs, one or more sets of biomedical outputs, one or more combined biomedical outputs, one or more sets of combined biomedical outputs. The set of biomedical outputs may include one or more biomedical outputs, one or more combined biomedical outputs, or a combination thereof. The combined biomedical outputs may include one or more biomedical outputs, one or more sets of biomedical outputs, one or more combined biomedical outputs, or a combination thereof. The set of combined biomedical outputs may include one or more biomedical outputs, one or more sets of biomedical outputs, one or more combined biomedical outputs, or a combination thereof. The one or more biomedical outputs, the one or more sets of biomedical outputs, the one or more combined biomedical outputs, the one or more sets of combined biomedical outputs may be based on or derived from one or more data, one or more datasets, one or more combined data, one or more combined datasets, one or more results, one or more sets of results, one or more combined results, one or more outputs, one or more sets of outputs, one or more combined outputs, one or more sets of combined outputs, or combinations thereof. The one or more biomedical outputs may include biomedical information of a subject. The biomedical information of the subject may predict, diagnose, and / or prognose one or more biomedical features. The one or more biomedical features may include severity of a disease or condition, genetic risk of a disease or condition, reproductive risk, genetic risk to a fetus, risk of adverse drug reactions, efficacy of drug therapy, prediction of optimal drug dosage, transplant tolerance, or combinations thereof.

[0176] Certain embodiments may include one or more biomedical reports. The methods, libraries, kits, and systems described herein may include generating one or more biomedical reports. The one or more biomedical reports may be based on or derived from one or more data, one or more datasets, one or more combined data, one or more combined datasets, one or more results, one or more sets of results, one or more combined results, one or more outputs, one or more sets of outputs, one or more combined outputs, one or more sets of combined outputs, one or more biomedical outputs, one or more sets of biomedical outputs, combined biomedical outputs, one or more sets of biomedical outputs, or combinations thereof. The biomedical report may predict, diagnose, and / or prognose one or more biomedical features. The one or more biomedical features may include extent of a disease or condition, genetic risk of a disease or condition, reproductive risk, genetic risk to a fetus, risk of adverse drug reactions, efficacy of drug therapy, prediction of optimal drug dosage, transplant tolerance, or combinations thereof.

[0177] Certain embodiments may also include the transmission of one or more data, information, results, outputs, reports, or combinations thereof. For example, data / information based on or derived from one or more assays is transmitted to another device and / or instrument. In another example, the data, results, outputs, biomedical outputs, biomedical reports, or combinations thereof are transmitted to another device and / or instrument. Information derived from an algorithm may also be transmitted to another device and / or instrument. Information based on the analysis of one or more databases may be transmitted to another device and / or instrument. The transmission of the data / information may include data / information transfer from a first source to a second source. The first source and the second source may be in the same approximate location (e.g., in the same room, building, block, campus). Alternatively, the first source and the second source may be in multiple locations (e.g., multiple cities, states, countries, continents, etc.). The data, results, outputs, biomedical outputs, biomedical reports may be communicated to a patient and / or a medical professional.

[0178] The transmission may be based on an analysis of one or more data, results, information, databases, outputs, reports, or combinations thereof. For example, the transmission of a second report is based on an analysis of a first report. Alternatively, the transmission of a report is based on an analysis of one or more data or results. The transmission may be based on receipt of one or more requests. For example, the transmission of a report may be based on receipt of a request from a user (e.g., a patient, a healthcare professional, an individual).

[0179] The transmission of data / information may include digital or analog transmission. Digital transmission may include the physical transmission of data (digital bit streams) over a point-to-point or point-to-multipoint communication channel. Examples of such channels include copper wire, fiber optics, wireless communication channels, and storage media. Data may be represented as an electromagnetic signal, such as a voltage, radio wave, microwave, or infrared signal.

[0180] Analog transmission may include the transmission of continuously varying analog signals. A message may be represented either as a series of pulses according to a line code (baseband transmission) or as a limited set of continuously varying waveforms using digital modulation techniques (passband transmission). Passband modulation and corresponding demodulation (also known as detection) may be performed by modem equipment. According to the most common definition of a digital signal, both baseband and passband signals representing bit streams are considered digital transmissions, while in other definitions only baseband signals are considered digital and passband transmission of digital data is considered a form of digital to analog conversion.

[0181] Certain embodiments may include one or more sample identifiers. The sample identifiers may include labels, bar codes, and other indicia that may be associated with one or more samples and / or subsets of nucleic acid molecules. Certain embodiments may include one or more processors, one or more memory locations, one or more computers, one or more monitors, one or more computer software, one or more algorithms for associating data, results, output, biomedical output, and / or biomedical reports with the samples.

[0182] Certain embodiments may include a processor for correlating expression levels of one or more nucleic acid molecules with a prognosis of disease outcome. Certain embodiments may include one or more of a variety of correlation techniques, including look-up tables, algorithms, multivariate models, and linear or non-linear combinations of formula models or algorithms. The expression levels may be converted into one or more likelihood scores reflecting the likelihood that the patient providing the sample will exhibit a particular disease outcome. The models and / or algorithms may be provided in a machine-readable format and may further specify a treatment for a patient or class of patients, if desired.

[0183] In some cases, the methods and systems described herein are used to generate output including detection and / or quantification of genomic DNA regions, such as regions containing DNA polymorphisms (e.g., germline variants or somatic variants). In some cases, the detection of one or more genomic regions is based on one or more algorithms, depending on the source of the data input or databases described elsewhere herein. Each of the one or more algorithms may be used to receive, combine, and generate data including detection of genomic regions (i.e., polymorphisms). In some embodiments, the methods and systems of the present invention may include detection of genomic regions based on one or more, two or more, three or more, four or more, five or more, six or more, seven or more, eight or more, nine or more, or ten or more algorithms. The algorithms may be machine learning algorithms, computer-implemented algorithms, machine-executed algorithms, automated algorithms, etc.

[0184] The resulting data for each nucleic acid sample can be analyzed using feature selection methods, including filter methods that assess the relevance of features by examining inherent properties of the data, wrapper methods that embed model hypotheses within the feature subset search, and embedded methods where the search for the optimal feature set is built into an algorithm or model.

[0185] In some cases, the detection of one or more genomic regions is based on one or more statistical models. Statistical models or filtering techniques useful in the methods of the present invention include: (1) parametric methods such as two-sample t-test, ANOVA analysis, Bayesian framework, and use of gamma distribution model; (2) model-free methods such as Wilcoxon rank sum test, inter-class and intra-class sum of squares test, rank product method, random pure quantification method, or use of TNoM, which includes setting a threshold point of difference in expression fold change between two data sets, and then finding the threshold point of each gene that minimizes the number of misclassifications; and (3) multivariate methods such as bivariate methods, correlation-based feature selection method (CFS), minimum redundancy maximum relevance method (MRMR), Markov blanket filter method, Markov model, hidden Markov model (HMM), and uncorrelated shrunken centroid (USC) method. In some cases, the hidden Markov model (HMM) is given an internal state, and the internal state is set according to the overall copy number of chromosomes of the first nucleic acid sample or the second nucleic acid sample. In one example, for diploid chromosomes, the internal states of the HMM can be homozygous deletion (zero copies locally), heterozygous deletion (one copy locally), normal (two copies locally), duplication (two or more copies), and reference gap (present as a state to distinguish gaps from homozygous deletions). In another example, for haploid chromosomes (e.g., male X or Y), the internal states of the HMIM can be homozygous deletion (zero copies locally), normal (two copies locally), duplication (two or more copies), and reference gap (present as a state to distinguish gaps from homozygous deletions). For example, for haploid chromosomes, there may be no available heterozygous deletion states. In another example, for trisomies and / or tetrasomies, the additional intermediate HMM states may have additional intermediate states, which may account for various CNV possibilities. In another embodiment, a hidden Markov model is used to filter the output by examining the insertion size of the reads measured near the breakpoints of the detected features.

[0186] Other models or algorithms useful in the methods of the present invention include sequential search methods, genetic algorithms, estimate of distribution algorithms, random forest algorithms, weight vectors of support vector machine algorithms, weights of logistic regression algorithms, etc. Bioinformatics. 2007 Oct 1;23(19):2507-17 provides an overview of the relative merits of the algorithms or models provided above for data analysis. Illustrative algorithms include, but are not limited to, methods that reduce the number of variables, such as principal component analysis algorithms, methods that directly handle a large number of variables, such as partial least squares, independent component analysis algorithms, statistical methods, and methods based on machine learning methods. Statistical methods include penalized logistic regression, predictive analysis of microarrays (PAM), methods based on reduced centroids, support vector machine analysis, and regularized linear discriminant analysis.

[0187] In some embodiments, HMM-based detection algorithms can detect large or substantially large CNVs "segmentally". In some cases, there may be small detection gaps along the length of the true CNV due to variations in coverage signal. In one example, a 1 megabase pair (Mbp) deletion may be detected as a few separate nominal detections with a small gap between them. To mitigate this, a merge operation may be used that identifies pairs of adjacent detections separated by a gap smaller than either of the two bracketed detections. The merge operation then measures the median coverage level within the gap. If the median coverage exceeds a predefined threshold, the two detections are merged into a single large detection that spans the two original detections (including the enclosed detection gap). In one example, the true feature spans both detections, and the gap is a statistical artifact. Using actual sequencing data of a sample known to have a large CNV, this merge operation can significantly increase fidelity with respect to the actual characteristics of the CNV.

[0188] The methods and systems provided herein may further include the use of feature selection algorithms as provided herein. In some embodiments of the invention, feature selection is provided using the LIMMA software package (Smyth, GK (2005). Limma: linear models for microarray data. In: Bioinformatics and Computational Biology Solutions using R and Bioconductor, R. Gentleman, V. Carey, S. Dudoit, R. Irizarry, W. Huber (eds.), Springer, New York, pages 397-420).

[0189] In some embodiments of the present invention, diagonal linear discriminant analysis, K-nearest neighbor algorithm, support vector machine (SVM) algorithm, linear support vector machine, random forest algorithm, probabilistic model-based method, or combinations thereof are provided to detect one or more genomic regions. In some embodiments, the identified markers that distinguish samples (e.g., disease vs. normal) or genomic regions (e.g., copy number variation vs. normal) are selected based on the statistical significance of the expression level difference between classes of interest. In some cases, the statistical significance is adjusted by applying Benjamini Hochberg or another correction for false discovery rate (FDR).

[0190] In some cases, the algorithms may be complemented with meta-analysis approaches such as those described in Fishel and Kaufman et al. 2007 Bioinformatics 23(13): 1599-606. In some cases, the algorithms may be complemented with meta-analysis approaches such as reproducibility analysis. In some cases, the reproducibility analysis selects markers that appear in at least one predicted expression product marker set.

[0191] Statistical evaluation of the detection of genomic regions provides one or more quantitative values ​​indicative of one or more of the following: likelihood of diagnostic accuracy; likelihood of injury, disease, condition, etc.; likelihood of a particular injury, disease, or condition; and likelihood of success of a particular therapeutic intervention. Thus, physicians, who likely have no training in genetics or molecular biology, do not need to understand the raw data. Rather, the data is presented directly to the physician in the form of quantitative values ​​to guide patient care. The results can be statistically evaluated using several methods known in the art, including, but not limited to, Student's t-test, two-tailed t-test, Pearson rank sum analysis, hidden Markov model analysis, analysis of QQ plots, principal component analysis, one-way analysis of variance, two-way analysis of variance, LIMMA, and the like.

[0192] F. Diseases and Conditions Certain embodiments may include predicting, diagnosing, and / or prognosing the extent or outcome of a disease or condition in a subject based on one or more biomedical outputs. Predicting, diagnosing, and / or prognosing the extent or outcome of a disease in a subject may include diagnosing a disease or condition, identifying a disease or condition, determining a stage of a disease or condition, assessing risk of a disease or condition, assessing risk of disease recurrence, assessing efficacy of a drug, assessing risk of adverse drug reactions, predicting optimal drug dosage, predicting drug resistance, or a combination thereof.

[0193] The sample disclosed herein may be from a subject suffering from cancer. The sample may include malignant tissue, benign tissue, or a mixture thereof. The cancer may be recurrent and / or refractory cancer. Examples of cancer include, but are not limited to, sarcoma, carcinoma, lymphoma, or leukemia. In some cases, a sample containing cancer tissue is obtained, but a matching normal sample is not obtained. In some cases, there is no matching normal sample available. In some cases, a matching normal sample is obtained (e.g., for the purpose of training and testing the model disclosed herein).

[0194] Sarcomas are cancers of bone, cartilage, fat, muscle, blood vessels, or other connective or supporting tissues. Sarcomas include, but are not limited to, bone cancer, fibrosarcoma, chondrosarcoma, Ewing's sarcoma, malignant hemangioendothelioma, malignant schwannoma, bilateral vestibular schwannoma, osteosarcoma, soft tissue sarcomas (e.g., alveolar soft part sarcoma, angiosarcoma, phyllodes cystosarcoma, dermatofibrosarcoma, desmoid tumor, epithelioid sarcoma, extraskeletal osteosarcoma, fibrosarcoma, hemangiopericytoma, angiosarcoma, Kaposi's sarcoma, leiomyosarcoma, liposarcoma, lymphangiosarcoma, lymphosarcoma, malignant fibrous histiocytoma, neurofibrosarcoma, rhabdomyosarcoma, and synovial sarcoma).

[0195] Carcinomas are cancers that arise from epithelial cells, which are cells that cover the body's surfaces, produce hormones, and make up glands. Non-limiting examples of carcinomas include breast cancer, pancreatic cancer, lung cancer, colon cancer, colorectal cancer, rectal cancer, kidney cancer, bladder cancer, stomach cancer, prostate cancer, liver cancer, ovarian cancer, brain cancer, vaginal cancer, vulvar cancer, uterine cancer, oral cancer, penile cancer, testicular cancer, esophageal cancer, skin cancer, fallopian tube cancer, head and neck cancer, gastrointestinal stromal cancer, adenocarcinoma, skin or intraocular melanoma, anal region cancer, small intestine cancer, endocrine system cancer, thyroid cancer, parathyroid cancer, adrenal gland cancer, urethral cancer, renal pelvis cancer, ureter cancer, endometrial cancer, cervical cancer, pituitary cancer, neoplasms of the central nervous system (CNS), primary CNS lymphoma, brain stem glioma, and spinal axis tumor. The cancer may be a skin cancer, such as a basal cell carcinoma, squamous cell carcinoma, melanoma, non-melanoma, or actinic (solar) keratosis.

[0196] The cancer may be lung cancer. Lung cancer may occur in the airways that branch off from the trachea and supply the lungs (bronchi) or the small air sacs of the lungs (alveoli). Lung cancer includes non-small cell lung cancer (NSCLC), small cell lung cancer, and mesothelioma. Examples of NSCLC include squamous cell carcinoma, adenocarcinoma, and large cell carcinoma. Mesothelioma may be a cancerous tumor of the lining of the lungs and chest cavity (pleura) or the abdominal lining (peritoneum). Mesothelioma may be caused by asbestos exposure. The cancer may be a brain tumor, such as glioblastoma.

[0197] The cancer may be a central nervous system (CNS) tumor. CNS tumors may be classified as gliomas or non-gliomas. Gliomas may be malignant gliomas, high-grade gliomas, diffuse intrinsic pontine gliomas. Examples of gliomas include astrocytomas, oligodendrogliomas (or a mixture of oligodendrogliomas and astrocytomas), and ependymomas. Astrocytomas include, but are not limited to, low-grade astrocytomas, anaplastic astrocytomas, glioblastoma multiforme, pilocytic astrocytomas, multiform xanthoastrocytomas, and subependymal giant cell astrocytomas. Oligodendrogliomas include low-grade oligodendrogliomas (or oligodendrogliomas) and anaplastic oligodendrogliomas. Non-gliomas include meningiomas, pituitary adenomas, primary CNS lymphomas, and medulloblastomas. The cancer may be meningiomas.

[0198] The leukemia can be acute lymphocytic leukemia, acute myeloid leukemia, chronic lymphocytic leukemia, or chronic myelogenous leukemia. Additional types of leukemia include hairy cell leukemia, chronic myelomonocytic leukemia, and juvenile myelomonocytic leukemia.

[0199] Lymphoma is a cancer of lymphocytes and can arise from either B or T lymphocytes. The two main types of lymphoma are Hodgkin's lymphoma, formerly known as Hodgkin's disease, and non-Hodgkin's lymphoma. Hodgkin's lymphoma is characterized by the presence of Reed-Sternberg cells. Non-Hodgkin's lymphoma is any lymphoma that is not Hodgkin's lymphoma. Non-Hodgkin's lymphoma can be indolent lymphoma and aggressive lymphoma. Non-Hodgkin's lymphomas include, but are not limited to, diffuse large B-cell lymphoma, follicular lymphoma, mucosa-associated lymphoid tissue lymphoma (MALT), small cell lymphocytic lymphoma, mantle cell lymphoma, Burkitt's lymphoma, mediastinal large B-cell lymphoma, Waldenstrom's macroglobulinemia, nodal marginal zone B-cell lymphoma (NMZL), splenic marginal zone lymphoma (SMZL), extranodal marginal zone B-cell lymphoma, intravascular large B-cell lymphoma, primary effusion lymphoma, and lymphomatoid granulomatosis.

[0200] Certain embodiments may include treating and / or preventing a disease or condition in a subject based on one or more biomedical outputs. The one or more biomedical outputs may recommend one or more therapies. The one or more biomedical outputs may suggest, select, prescribe, recommend, or otherwise determine a course of treatment and / or prevention for a disease or condition. The one or more biomedical outputs may recommend modifying or continuing one or more therapies. Modifying one or more therapies may include administering, initiating, reducing, increasing, and / or terminating one or more therapies. The one or more therapies may include an anti-cancer therapy, an anti-viral therapy, an anti-bacterial therapy, an anti-fungal therapy, an immunosuppressive therapy, or a combination thereof. The one or more therapies may treat, alleviate, or prevent one or more diseases or indications.

[0201] Examples of anti-cancer therapies include, but are not limited to, surgery, chemotherapy, radiation therapy, immunotherapy / biological therapy, photodynamic therapy. Anti-cancer therapies may include chemotherapeutic agents, monoclonal antibodies (e.g., rituximab, trastuzumab), cancer vaccines (e.g., therapeutic vaccines, prophylactic vaccines), gene therapy, or a combination thereof.

[0202] G. Systems, Kits, and Libraries The disclosed method may be implemented by a system, a kit, a library, or a combination thereof. The disclosed method may include one or more systems. The disclosed system may be implemented via a kit, a library, or both. The system may include one or more components for implementing any of the methods or steps of a particular embodiment. For example, the system may include one or more kits, devices, libraries, or combinations thereof. The system may include one or more sequencers, processors, memory locations, computers, computer systems, or combinations thereof. The system may include a transmission device.

[0203] The kit may include various reagents for performing various operations disclosed herein, including sample processing and / or analytical operations. The kit may include instructions for performing at least some of the operations disclosed herein. The kit may include one or more capture probes, one or more beads, one or more labels, one or more linkers, one or more devices, one or more reagents, one or more buffers, one or more samples, one or more databases, or a combination thereof.

[0204] The library may include one or more capture probes. The library may include a subset of one or more nucleic acid molecules. The library may include one or more databases. The library may be created or generated from any of the methods, kits, or systems disclosed herein. A database library may be created from one or more databases. A method of creating one or more libraries may include (a) aggregating information from one or more databases to create an aggregated data set, (b) analyzing the aggregated data set, and (c) creating one or more database libraries from the aggregated data set.

[0205] VI. Computing Environment FIG. 10 illustrates an example of a computer system 1000 for implementing some of the embodiments disclosed herein. The computer system 1000 may have a distributed architecture where some components (e.g., memory and processor) are part of an end-user device and some other similar components (e.g., memory and processor) are part of a computer server. The computer system 1000 includes at least a processor 1002, a memory 1004, a storage device 1006, input / output (I / O) peripherals 1008, communication peripherals 1010, and an interface bus 1012. The interface bus 1012 is configured to communicate, transmit, and transfer data, control, and commands between various components of the computer system 1000. The processor 1002 may include one or more processing units, such as a CPU, a GPU, a TPU, a systolic array, or a SIMD processor. The memory 1004 and storage 1006 include computer-readable storage media such as RAM, ROM, electrically erasable programmable read-only memory (EEPROM), hard drives, CD-ROM, optical storage, magnetic storage, electronic non-volatile computer storage (e.g., Flash), and other tangible storage media. Any such computer-readable storage media may be configured to store instructions or program code embodying aspects of the present disclosure. The memory 1004 and storage 1006 also include computer-readable signal media. Computer-readable signal media includes a propagated data signal having computer-readable program code embodied therein. Such a propagated signal may take any of a variety of forms, including, but not limited to, electromagnetic, optical, or any combination thereof. Computer-readable signal media includes any computer-readable medium that is not a computer-readable storage medium and may communicate, propagate, or transmit a program for use in connection with the computer system 1000.

[0206] Additionally, the memory 1004 includes an operating system, programs, and applications. The processor 1002 is configured to execute stored instructions and includes, for example, logic processing units, microprocessors, digital signal processors, and other processors. The memory 1004 and / or the processor 1002 can be virtualized and hosted within another computing system, for example, a cloud network or a data center. The I / O peripherals 1008 include user interfaces such as keyboards, screens (e.g., touch screens), microphones, speakers, and other input / output devices, as well as computing components such as graphical processing units, serial ports, parallel ports, universal serial buses, and other input / output peripherals. The I / O peripherals 1008 are connected to the processor 1002 via any ports coupled to an interface bus 1012. The communication peripherals 1010 are configured to facilitate communication between the computer system 1000 and other computing devices over a communication network and include, for example, network interface controllers, modems, wireless and wired interface cards, antennas, and other communication peripherals.

[0207] Although the subject matter of the present invention has been described in detail with respect to certain embodiments thereof, it will be understood that those skilled in the art, upon gaining the foregoing understanding, may easily make modifications, variations, and equivalents of such embodiments. It is therefore to be understood that the present disclosure is presented for purposes of example and not limitation, and is not intended to exclude the inclusion of such modifications, variations, and / or additions to the subject matter of the present invention as would be readily apparent to one skilled in the art. Indeed, the methods and systems described herein may be embodied in a variety of other forms, and further, various omissions, substitutions, and changes in the form of the methods and systems described herein may be made without departing from the spirit of the disclosure. The accompanying claims and their equivalents are intended to cover such forms or modifications as fall within the scope and spirit of the present disclosure.

[0208] Unless otherwise indicated, use of terms such as "processing," "computing," "calculating," "determining," and "identifying" are understood throughout the description herein to refer to the actions or processes of a computing device, such as one or more computers or similar electronic computing devices, that manipulate or transform data represented as physical electronic or magnetic quantities within the memory, registers, or other information storage, transmission, or display devices of a computing platform.

[0209] The system or systems described herein are not limited to any particular hardware architecture or configuration. A computing device may include any suitable arrangement of components that provide results conditioned on one or more inputs. Suitable computing devices include general-purpose microprocessor-based computing systems that access stored software that programs or configures the computing system, ranging from general-purpose computing devices to specialized computing devices that implement one or more embodiments of the inventive subject matter. Any suitable programming, scripting, or other type of language or combination of languages ​​may be used to implement the teachings contained herein in software used to program or configure a computing device.

[0210] The method embodiments disclosed herein may be performed in operation of such a computing device. The order of the blocks illustrated in the above examples may be changed, e.g., the blocks may be reordered, combined, and / or divided into sub-blocks. Certain blocks or processes may be performed in parallel.

[0211] Conditional language used herein, such as "can," "could," "might," "may," "for example," and the like, is intended to generally convey that certain examples include certain features, elements, and / or steps, but not other examples, unless otherwise specified or otherwise understood within the context in which it is used. Thus, such conditional language is not intended to generally imply that features, elements, and / or steps are required for one or more examples, or that one or more examples necessarily include logic for determining, with or without author input or direction, that those features, elements, and / or steps are included or performed in a particular example.

[0212] Terms such as "comprising," "including," "having," and the like are synonymous and are used inclusively in an open-ended manner and do not exclude additional elements, features, actions, operations, and the like. Additionally, the term "or" is used in its inclusive (not exclusive) sense, so that, for example, when used to relate a list of elements, the term "or" means one, some, or all of the elements in the list. The use of "adapted" or "configured" herein means open and inclusive language that does not exclude devices adapted or configured to perform additional tasks or steps. Additionally, the use of "based on" means open and inclusive in that a process, step, calculation, or other action that is "based on" one or more recited conditions or values ​​may in fact be based on additional conditions or values ​​beyond the recited values. Similarly, the use of "based at least in part on" means open and inclusive in that a process, step, calculation, or other action that is "based at least in part on" one or more recited conditions or values ​​may in fact be based on additional conditions other than those recited or values ​​beyond the recited values. The headings, lists, and numbering contained herein are for ease of explanation and are not limiting.

[0213] The various features and processes described above may be used independently of one another or may be combined in various ways. All possible combinations and subcombinations are intended to fall within the scope of the present disclosure. Furthermore, in some implementations, certain method or process blocks may be omitted. The methods and processes described herein are also not limited to a particular sequence, and the blocks or states associated therewith may be performed in other sequences as appropriate. For example, the described blocks or states may be performed in an order other than the order specifically disclosed, or multiple blocks or states may be combined into a single block or state. The examples of blocks or states may be performed sequentially, in parallel, or in other ways. Blocks or states may be added to or deleted from the disclosed examples. Similarly, the example systems and components described herein may be configured differently than as described. For example, elements may be added, deleted, or rearranged compared to the disclosed examples.

Claims

1. 1. A method for determining tumor purity, comprising: obtaining nucleic acid sequence data representing a plurality of nucleic acid molecules of a tumor sample from a subject; aligning the nucleic acid sequence data to a reference genome; identifying a set of genomic regions based on the aligned nucleic acid sequence data, wherein each genomic region of the set of genomic regions comprises one or more nucleotide sequence variants relative to a corresponding genomic region of the reference genome; determining the B allele frequency of each genomic region of said set of genomic regions; determining a B allele frequency distribution for the tumor sample based on the B allele frequencies of the set of genomic regions; processing the B allele frequency distribution using a trained machine learning model to estimate a metric that identifies the tumor purity of the tumor sample; and and outputting the metric, wherein the trained machine learning model has been trained with a training dataset generated from sequencing datasets from pure tumor samples, pure normal samples, and a dilution series of tumor cells among normal cells.

2. 2. The method of claim 1, wherein the nucleic acid sequence data is whole exome sequencing data.

3. The method of claim 1 , wherein the nucleic acid sequence data is whole genome sequencing data.

4. obtaining the tumor sample from the subject; and sequencing the plurality of nucleic acid molecules of the tumor sample to generate the nucleic acid sequence data; The method of claim 1 further comprising:

5. 5. The method of claim 4, further comprising isolating the plurality of nucleic acid molecules prior to sequencing.

6. The method of claim 1 , wherein the B allele frequency distribution is normalized.

7. The method of claim 1 , further comprising outputting a report including information identifying the B allele frequency distribution.

8. 10. The method of claim 1, further comprising outputting a report including a metric identifying the estimated tumor purity.

9. The method of claim 8 , wherein the report further comprises information identifying at least one biomarker.

10. The method of claim 8 , wherein the report further comprises information identifying at least one prognostic marker.

11. 10. The method of claim 8, wherein the report includes information identifying predicted somatic variants.

12. The method of claim 8 , wherein the report includes a treatment recommendation.

13. The method of claim 1 , wherein the tumor sample is from a human subject.

14. 13. The method of claim 12, wherein the treatment recommendation comprises a recommendation to not administer treatment to the subject.

Images