Techniques for the projection of single sample expression onto expression cohorts sequenced using alternative protocols

Abstract

Aspects of the present disclosure relate to a method for improving the compatibility of nucleic acid sequencing data obtained using different techniques. The present disclosure is based, in part, on a method for mapping expression levels of genes expressed in a biological sample and obtained from a subject using a first protocol to expression levels that would be determined by a second protocol if the biological sample was processed using the second protocol instead of the first protocol.

Description

[Technical field]

[0001] Related Applications This application claims the benefit under 35 U.S.C. § 119(e) of the filing date of U.S. Provisional Patent Application No. 63 / 190,171, filed May 18, 2021, the entire contents of which are incorporated herein by reference. [Background technology]

[0002] Gene expression profiling (GEP) is a powerful tool that is widely used in tumor research. GEP uses technologies such as NGS and microarrays to simultaneously evaluate the expression levels of multiple genes. Each expression level measurement is typically evaluated against a cohort of samples that have been sequenced using the same method to understand whether the sample has a high or low expression level value. Summary of the Invention [Means for solving the problem]

[0003] Aspects of the present disclosure relate to a method for improving the compatibility of nucleic acid sequencing data obtained using different techniques. The present disclosure is based, in part, on a method for mapping RNA expression levels of genes expressed in a biological sample obtained from a subject using a first protocol to RNA expression levels that would be determined by a second protocol if the biological sample was processed using the second protocol instead of the first protocol.

[0004] Thus, in some aspects, the disclosure provides a method for mapping RNA expression levels of genes expressed in a biological sample obtained from a subject using a first protocol to RNA expression levels that would have been determined by a second protocol if the biological sample was processed using a second protocol instead of the first protocol, the method comprising: obtaining first RNA expression data for a set of genes expressed in a biological sample obtained from a subject, the first RNA expression data being indicative of a first RNA expression level (e.g., comprising a first RNA expression level) of genes in the set of genes, the first RNA expression data comprising a first RNA expression level, the first RNA expression data being pre-determined by processing the biological sample using the first protocol; and obtaining a second RNA expression level for a set of genes expressed in a biological sample obtained from a subject, the set of genes being indicative of a first RNA expression level (e.g., comprising a first RNA expression level), the second RNA expression level being indicative of a first RNA expression level, the second RNA expression level being pre-determined by processing the biological sample using the first protocol; The present invention provides a method for determining a first RNA expression level for a gene in a set of genes by applying the first transformation to the first set of RNA expression levels for the gene in the set of genes, the method comprising: obtaining, for a first gene in the set of genes, from among the first RNA expression levels, a first set of RNA expression levels including a first RNA expression level for the first gene and zero, one, or more first RNA expression levels for zero, one, or more other genes in the set of genes related to the first gene; obtaining a first transformation from the RNA expression levels of the one or more genes determined by the first protocol to estimate an RNA expression level for the first gene that would have been determined according to a second protocol; and determining a second RNA expression level for the first gene by applying the first transformation to the first set of RNA expression levels for inclusion in the second RNA expression level.

[0005] In some aspects, the disclosure provides a system that includes at least one computer hardware processor; and at least one computer readable storage medium storing processor-executable instructions that, when executed by the at least one computer hardware processor, cause the at least one computer hardware processor to execute a method for mapping RNA expression levels of genes expressed in a biological sample obtained from a subject using a first protocol to RNA expression levels that would have been determined by a second protocol if the biological sample were processed using the second protocol instead of the first protocol, the method comprising: obtaining first RNA expression data for a set of genes expressed in a biological sample obtained from the subject, the first RNA expression data being indicative of a first RNA expression level (e.g., comprising a first RNA expression level) of genes in the set of genes, the first RNA expression data being previously determined by processing the biological sample using the first protocol; and the second RNA expression level indicates an RNA expression level that would have been determined by a second protocol if the biological sample had been processed using the second protocol instead of the first protocol, the second protocol differing from the first protocol; using at least one computer hardware processor to perform: mapping a first RNA expression level of a gene in the set of genes to a second RNA expression level of a gene in the set of genes, the mapping comprising: obtaining, for a first gene in the set of genes, from among the first RNA expression levels a first set of RNA expression levels comprising a first RNA expression level for the first gene and zero, one or more first RNA expression levels for zero, one or more other genes in the set of genes related to the first gene; obtaining a first transformation from the RNA expression level of the one or more genes determined by the first protocol to estimate an RNA expression level for the first gene that would have been determined according to the second protocol;and determining a second RNA expression level for the first gene by applying a first transformation to the first set of RNA expression levels for inclusion in the second RNA expression level;

[0006] In some embodiments, the processor-executable instructions, when executed by at least one computer hardware processor, cause the at least one computer hardware processor to perform the methods described herein.

[0007] In some aspects, the disclosure provides at least one computer readable storage medium storing processor-executable instructions that, when executed by at least one computer hardware processor, cause the at least one computer hardware processor to execute a method for mapping RNA expression levels of genes expressed in a biological sample obtained from a subject using a first protocol to RNA expression levels that would be determined by a second protocol if the biological sample were processed using the second protocol instead of the first protocol, the method comprising: obtaining first RNA expression data for a set of genes expressed in a biological sample obtained from the subject, the first RNA expression data being indicative of a first RNA expression level (e.g., comprising a first RNA expression level) of genes in the set of genes, the first RNA expression data being previously determined by processing the biological sample using the first protocol; and the first protocol indicates an RNA expression level that would have been determined by a second protocol if the biological sample had been processed using a second protocol instead of the first protocol, the second protocol differing from the first protocol, using at least one computer hardware processor to perform mapping a first RNA expression level of a gene in the set of genes to a second RNA expression level of a gene in the set of genes, the mapping comprising, for a first gene in the set of genes, from among the first RNA expression levels, a first set of RNA expression levels including a first RNA expression level for the first gene and zero, one, or more first RNA expression levels for zero, one, or more other genes in the set of genes related to the first gene; obtaining a first transformation from the RNA expression levels of the one or more genes determined by the first protocol to estimate an RNA expression level for the first gene that would have been determined according to the second protocol;and determining a second RNA expression level for the first gene by applying a first transformation to the first set of RNA expression levels for inclusion in the second RNA expression level.

[0008] In some embodiments, the method further comprises identifying, from among the multiple cohorts, a cohort associated with the subject using the second RNA expression level.

[0009] In some embodiments, the set of genes includes a second gene and a second set of genes associated with the second gene, and the mapping includes obtaining, from among the first RNA expression levels, a second set of RNA expression levels including a first RNA expression level for the second gene and RNA expression levels for genes in the second set of genes associated with the second gene; obtaining a second transformation for estimating an RNA expression level for the second gene determined according to a second protocol from the RNA expression levels of the one or more genes determined by the first protocol, the second transformation being different from the first transformation; and determining a second RNA expression level for the second gene by applying the second transformation to the second set of RNA expression levels for inclusion in the second RNA expression level.

[0010] In some embodiments, the set of genes includes one or more additional genes and a further set of genes associated with the one or more additional genes, and the mapping comprises: obtaining a set of RNA expression levels from the first RNA expression level, the set including an RNA expression level for at least a portion of each of the one or more additional genes and an RNA expression level for at least a portion of the genes of the further set of genes associated with the one or more additional genes; obtaining a respective transformation for estimating an RNA expression level for each of the one or more additional genes determined according to a second protocol; and determining a second RNA expression level for each of at least a portion of the additional genes of the subset by applying the second transformation to the first set of RNA expression levels for inclusion in the second RNA expression level. In some embodiments, the set of RNA expression levels includes an RNA expression level for at least a portion of each of the one or more additional genes and an RNA expression level for at least a portion of the genes of the further set of genes associated with the one or more additional genes.

[0011] In some embodiments, the method includes, prior to the mapping, determining, for each gene of at least a subset of the set of genes, a respective transformation for estimating, from the RNA expression levels of one or more genes of the subset determined by the first protocol, an RNA expression level for each gene of the subset that would have been determined according to the second protocol.

[0012] In some embodiments, the transformation is a linear transformation and the determination of the first transformation is performed using a regularized linear regression technique using the training data, hi some embodiments, the transformation is a non-linear transformation and the determination of the first transformation is performed using a non-linear regression technique using the training data.

[0013] In some embodiments, the training data comprises a plurality of paired values ​​of RNA expression levels for each of at least a portion of the set of genes, each pair of values ​​in the plurality of paired values ​​comprising an RNA expression level determined by applying a first protocol to a particular biological sample and another RNA expression level determined by applying a second protocol to the particular biological sample.

[0014] In some embodiments, obtaining the first set of expression levels comprises obtaining a first expression level for a first gene and zero other RNA expression levels.

[0015] In some embodiments, obtaining the first set of RNA expression levels comprises identifying one or more other genes associated with the first gene. In some embodiments, the identification is performed using Pearson correlation.

[0016] In some embodiments, the plurality of other genes in the set of genes comprises between 2 and 100 genes related to the first gene.

[0017] In some embodiments, the biological sample comprises a blood sample or a tissue sample. In some embodiments, the tissue sample comprises tumor tissue.

[0018] In some embodiments, the subject is a mammal, hi some embodiments, the subject is a human.

[0019] In some embodiments, the first RNA expression data and the second RNA expression data comprise normalized RNA expression levels. In some embodiments, the normalized RNA expression levels are normalized to transcripts per million (TPM).

[0020] In embodiments, the first protocol and the second protocol each include one or more sample processing steps and sequencing steps, and the first protocol includes a sample processing step and / or a sequencing step that does not form part of the second protocol. In some embodiments, the first protocol includes preserving the biological sample by formalin fixation and paraffin embedding (FFPE) technique. In some embodiments, the first protocol further includes performing exome capture (EC) RNA sequencing on the biological sample preserved by FFPE.

[0021] In some embodiments, the second protocol comprises preserving the biological sample by fresh freezing (FF) technique. In some embodiments, the second protocol comprises performing polyA RNA sequencing on the biological sample preserved by FF.

[0022] In some embodiments, the method further comprises generating first RNA expression data by applying the first protocol to the biological sample.

[0023] In some embodiments, identifying the cohort comprises correlating the second RNA expression level with the RNA expression level of a particular cohort of the plurality of cohorts; and identifying the subject as a member of the particular cohort to which the second RNA expression level is associated.

[0024] In some embodiments, the method further comprises using the second RNA expression level to select a cancer therapeutic agent for the subject.In some embodiments, selecting the cancer therapeutic agent comprises: determining the RNA expression level of a plurality of gene groups using the second RNA expression level, wherein the RNA expression level of the plurality of gene groups comprises the RNA expression level of the gene group for each gene group in the set of gene groups, and the set of gene groups comprises at least one gene group associated with cancer malignancy and at least one gene group associated with cancer microenvironment; and using the determined RNA expression level of the gene group to select a cancer therapeutic agent.

[0025] In some embodiments, the method further comprises administering to the subject a selected cancer therapeutic agent. [Brief description of the drawings]

[0026] [Figure 1A] 1 is a schematic diagram showing that RNA expression data obtained from a single biological sample using a first protocol (e.g., exome capture (EC) RNA sequencing) is not comparable to reference RNA expression data obtained from a sample obtained using a different protocol (e.g., polyA RNA sequencing). [Figure 1B] 1 is a schematic diagram showing that methods according to some embodiments of the technology described herein (e.g., single sample mapping) can be applied to RNA expression data obtained from a single biological sample using a first protocol (e.g., exome capture (EC) RNA sequencing) to generate RNA expression data for the biological sample that can be compared to reference RNA expression data obtained from a sample obtained using a different protocol (e.g., polyA RNA sequencing). [Figure 2A] 1 is a diagram depicting a single gene linear mapping technique according to some embodiments of the technology described herein. [Figure 2B] 1 is a diagram depicting a single gene general mapping technique according to some embodiments of the technology described herein. [Figure 2C] 1 is a diagram depicting a multi-gene linear mapping technique according to some embodiments of the technology described herein. [Figure 2D] 1 is a diagram depicting a multi-gene general mapping technique according to some embodiments of the techniques described herein. [Diagram 3]FIG. 3 depicts a flowchart of an exemplary process 300 for mapping RNA expression levels of genes expressed in a biological sample obtained from a subject using a first protocol to the RNA expression levels that would have been determined by a second protocol if the biological sample had been processed using the second protocol instead of the first protocol, in accordance with some embodiments of the technology described herein. [Figure 4] FIG. 1 depicts a flowchart of an exemplary process for mapping a first RNA expression level obtained from a subject using a first protocol to a second RNA expression level that would have been determined by a second protocol if the biological sample was processed using the second protocol instead of the first protocol, according to some embodiments of the technology described herein. [Diagram 5] FIG. 1 shows the number of sample pairs per diagnosis in the MER500 dataset. [Figure 6] FIG. 13 shows the principal component analysis (PCA) projection of expression of 320 paired RNA-seq samples per protocol in the MET500 cohort. [Figure 7] FIG. 1 shows expression (log2+1) correlations for a representative set of cancer or immune system genes, with exome capture (EC) values ​​plotted on the x-axis and polyA values ​​plotted on the y-axis. [Figure 8] FIG. 13 shows UMAP projection of batch effect-effective correlation of polyA samples with predicted samples mixed to preserve cancer-specific groupings. [Figure 9] FIG. 1 shows concordance correlation values ​​in biologically meaningful gene (BMG) space before and after correlation according to methods described in some embodiments of the technology described herein. [Figure 10] FIG. 13 shows microenvironment gene signature concordance correlation coefficient (CCC) values ​​for paired polyA and EC samples before and after correlation. [Figure 11]FIG. 1 shows the difference in CCC values ​​for each single sample gene set enrichment assay (ssGSEA) process. The correlation values ​​before correlation were subtracted from the correlation values ​​after correlation. The dotted line indicates the difference equal to 0. [Figure 12] FIG. 1 shows CCC values ​​for a representative analysis process before and after correlation of expression values. [Figure 13] Figure 1 shows the percentage of polyA vs. EC-predicted CD4+ T cell RNA (before re-normalization using the RNA per cell type factor) before (left) and after (right) correlation. The line represents y=x. [Figure 14] Figure 1 shows the Pearson correlation of expression values ​​for CXCR6 vs. CCR5. Efficiency of expression correlation for the CXCR6 gene: single gene vs. multigene technique (measured in CCC). [Figure 15] FIG. 13 shows CCC values ​​in BMG space before and after correlation with two developed "single gene" and "multi-gene" techniques, according to some embodiments of the technology described herein. [Figure 16] FIG. 1 shows the amount of variance due to each of the 20 principal components (PCs) of the pooled polyA and EC expression data. [Figure 17A] FIG. 1 shows the performance of the PCA method on the training set, excluding the first and second PCs. [Figure 17B] FIG. 1 shows the performance of the PCA method on the training set, excluding the third and fifth PCs. [Figure 18A] FIG. 13 shows the performance of the PCA method on a hold-out set, excluding the first and second PCs. [Figure 18B] FIG. 13 shows the performance of the PCA method on a holdout set, excluding the third and fifth PCs. [Figure 19] 1 is a diagram depicting the workflow for analysis based on the Multiple Neighborhood Neighborhood (MNN) transform. [Figure 20]Representative data for PCA on hold-out and MNN transformed data showing batch effects for paired samples sequenced using polyA RNA-seq vs. EC. "Original" refers to hold-out expression data before correlation. [Figure 21] FIG. 13 shows concordance correlation values ​​in BMG space before and after correlation using MNN compared to single gene sample mapping methods according to some embodiments of the technology described herein. [Figure 22] FIG. 13 shows concordance correlation values ​​in BMG space before and after correlation using ComBat compared to single gene sample mapping methods according to some embodiments of the technology described herein. [Diagram 23] FIG. 13 shows PCA on hold-out data showing batch effect after correlation of EC expression with ComBat. [Figure 24] FIG. 1 shows representative data for the performance of methods according to some embodiments of the technology described herein and other batch correction methods on four predefined gene groups. CCC values ​​are divided into three intervals. [Figure 25A] PCA on training data showing batch effect for paired samples sequenced using polyA RNA-seq vs. EC. The top plot is colored by protocol and the bottom plot is colored by sample type. [Figure 25B] FIG. 13 shows PCA on training data where different sample types separately indicate the presence of batch effects between protocols. [Figure 26] FIG. 1 shows PCA on validation data before correction for batch effect, with the top plot shaded by protocol and the bottom plot shaded by sample origin. [Figure 27] PCA on validation data after correction showing no batch effect. The top plot is shaded by protocol, the middle plot is shaded by sample origin, and the bottom plot is shaded by sample type. Points from the same sample are grouped together. [Figure 28] Figure 1 shows gene expression correlation between FF-polyA and FFPE-EC_V7 on the same sample. CCC values ​​are shown in the caption. [Figure 29] FIG. 1 shows representative data for corrected intra-sample correlations, the average means the intra-sample correlation is about 0.95. [Diagram 30] FIG. 1 shows the CCC distribution of BMG before correction, after correction using single-gene ElasticNetCV technique, and after correction using multi-gene CV technique. [Diagram 31] 1 illustrates the performance of methods according to some embodiments of the technology described herein on laboratory data, where the CCC values ​​are divided into three intervals. [Diagram 32] FIG. 32 illustrates an exemplary process 3200 for processing sequencing data to obtain RNA expression data from the sequencing data. [Diagram 33] FIG. 1 depicts an example implementation of a computer system that can be used in association with some embodiments of the technology described herein. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0027] Embodiments of the present disclosure relate to methods for improving the compatibility of nucleic acid sequencing data obtained using different protocols, e.g., RNA sequencing data obtained from samples prepared according to different storage, nucleic acid extraction, and / or nucleic acid sequencing techniques.

[0028] Significant variation in absolute expression value of a gene within a single biological sample may be caused by one or more differences in the protocol used to derive the absolute expression value (e.g., differences in storage, extraction, and / or nucleic acid sequencing techniques). Even when using the same protocol, significant variation in absolute expression value of a gene may be observed between samples that are not processed together or exactly the same (e.g., using different batches of reagents, different operators, different conditions, etc.). This variability may be referred to as batch effect in that it affects (affects) multiple samples that are processed using the same protocol (as a batch).

[0029] There are existing techniques to reduce the impact of such batch effects on genomic data. However, such techniques are only applicable in the context of reducing batch effects between samples across large cohorts. This is a major problem because such techniques cannot be applied to correct for batch effects when comparing individual samples with a reference cohort that contains multiple samples (single sample batch effect setting), and can only be used when comparing two cohorts that each contain several samples (multiple cohort batch effect setting).

[0030] This limitation of conventional techniques for correcting for batch effects in gene expression levels (e.g., RNA expression levels) is particularly problematic in current precision medicine applications. Many precision medicine applications include identifying biomarkers derived from sequencing data obtained from a subject (e.g., a subject who has cancer, is suspected to have cancer, or is at risk of having cancer), identifying a cohort of subjects by comparing the subject's biomarkers with the biomarkers of others in each of a plurality of cohorts, and performing diagnostic, prognostic, and / or therapeutic activities based on the identified cohort. Often, the biomarkers used are gene expression levels (e.g., RNA expression levels) themselves or are derived from gene expression levels (e.g., RNA expression levels). If the biomarkers of a subject depend on the gene expression levels (e.g., RNA expression levels) obtained using one protocol, and the biomarkers of a subject in the tested cohort depend on the gene expression levels (e.g., RNA expression levels) obtained using a different protocol, batch effects may make the comparison of biomarkers between subjects and cohorts inappropriate, inaccurate, and / or meaningless. Inappropriate diagnostic, prognostic, and / or treatment actions may result from such comparisons.

[0031] The following are specific examples of the situation. Biological samples are typically preserved and stored as fresh frozen (FF) or formalin-fixed paraffin-embedded (FFPE) samples. FF storage is uncommon in clinical practice because it requires the purchase and maintenance of expensive freezers. Nucleic acids are typically better preserved in FF samples, allowing for high-quality sequencing output. FFPE samples, on the other hand, are often used for routine pathology testing and are the primary method for clinical sample storage. However, the fixation step of FFPE storage induces changes in the nucleic acids. For example, FFPE processing physically crosslinks the nucleic acids and proteins in the sample, degrading long molecules into smaller fragments, resulting in challenges for downstream RNA extraction and sequencing. Furthermore, while fresh frozen samples can typically be sequenced using any of several different nucleic acid sequencing techniques (e.g., polyA RNA sequencing, exome capture RNA sequencing, etc.), samples prepared by FFPE are not suitable for polyA sequencing techniques because RNA derived from FFPE material is often degraded to small sizes and may lack polyA tails.

[0032] Continuing with this example, Figure 1A illustrates a challenge to the art of nucleic acid sequencing caused by the inapplicability of conventional techniques to address batch effect issues in a single sample setting. In Figure 1A, expression data (e.g., RNA expression data), 102, obtained from a single biological sample using a first protocol (e.g., FFPE preparation followed by exome capture (EC) RNA sequencing), is not comparable to reference expression data (e.g., reference RNA expression data for a patient cohort), 104, obtained from a sample obtained using a different protocol (e.g., FF preparation followed by polyA RNA sequencing). For example, The Cancer Genome Atlas (TCGA) has established a database of annotation-rich polyA RNA sequenced samples derived from FF tissues for over 30 cancer types, which is a valuable resource of sequencing data that can potentially be utilized as comparative gene expression profiling (GEP) cohorts (e.g., Figure 1A, 104). In contrast, samples obtained from cancer patients in the clinic almost exclusively contain tissues preserved using formalin-fixed paraffin-embedded (FFPE) histology (e.g., Figure 1A, 102). Because these patient samples cannot be sequenced using polyA sequencing, GEP is performed using exome capture (EC) RNA-seq protocols. However, EC protocols often differ and rely on custom-made gene panels; therefore, patient samples and cohorts are often sequenced using different protocols and panels.

[0033] As mentioned above, there are no available conventional techniques for generating gene expression data (e.g., RNA expression data) from single biological samples sequenced using exome capture technology that are compatible, and therefore meaningfully compatible, with polyA RNA-seq data. Thus, a large cohort of patient data (e.g., TCGA data) obtained by polyA RNA-seq of cancer research subjects may be of little use to physicians who need to analyze expression data obtained from FFPE patient samples sequenced by EC. Thus, the lack of compatibility between sequencing data on FF-stored samples at the single-sample level and sequencing data on FFPE-stored samples has a negative impact on the quality of bioinformatics analysis of patient samples and the application of cancer research exploration to clinical settings.

[0034] Thus, the inventors have developed statistical techniques to map RNA expression levels of genes expressed in a biological sample obtained from a subject using a first protocol to the RNA expression levels that would have been determined by a second protocol if the biological sample had been processed using the second protocol instead of the first protocol.

[0035] In some embodiments, the mapping can be performed on a gene-by-gene basis, such that each particular gene is associated with a respective mapping that is used to estimate the RNA expression level of one or more genes determined by applying a first protocol to the biological sample, and the RNA expression level of that particular gene determined by processing the biological sample using a second protocol instead. In some embodiments, the mapping can be a linear mapping (e.g., a linear transformation), and its exact value can be estimated from the training data using a linear regression technique (e.g., linear regression, least absolute shrinkage and selection operator (LASSO) regression, ridge regression, ElasticNet regression, or any other suitable regression or regularized regression technique), as described herein. Using the application of statistical techniques developed by the inventors, gene expression data (e.g., RNA expression data) of a biological sample can be made compatible with gene expression data obtained by other sample preparation or sequencing techniques, allowing direct single-sample comparison.

[0036] In particular, the above problems with respect to Figure 1A can be addressed by the technology developed by the present inventors. As shown in Figure 1B, an embodiment of the technology described herein can be implemented as part of a software module (e.g., shown as a "single sample mapping" software module, 106 in Figure 1B) that can be applied to RNA expression data, 102, obtained from a single biological sample using a first protocol (e.g., exome capture (EC) RNA sequencing) to generate RNA expression data (Figure 1B, 108) of the biological sample that can be compared with reference RNA expression data (e.g., TCGA data obtained by polyA RNA sequencing, Figure 1B, 104) obtained from a sample obtained using a different protocol.

[0037] Accordingly, some embodiments are computer-implemented methods for identifying a subject (e.g., a mammal, e.g., a human) as a member of a cohort, the method comprising: (A) obtaining first RNA expression data for a set of genes expressed in a biological sample (e.g., blood, tissue, tumor tissue) obtained from the subject, where the first RNA expression data indicates a first RNA expression level of genes in the set of genes, the first RNA expression data being previously determined by processing the biological sample using a first protocol; (B) mapping the first RNA expression level of genes in the set of genes to a second RNA expression level of genes in the set of genes, where the second RNA expression level is indicative of an RNA expression level that would have been determined by a second protocol, different from the first protocol, if the biological sample had been processed using the second protocol instead of the first protocol; and (C) mapping a plurality of cohorts (e.g., a plurality of cohorts of subjects, a plurality of cohorts of samples). and (iii) determining a second RNA expression level for the first gene by applying the first transformation to the first set of RNA expression levels for inclusion in the second RNA expression level.

[0038] A plurality of genes may have their RNA expression levels mapped from a "first protocol" value (actually measured) to a projected "second protocol value." Thus, in some embodiments, the set of genes includes a second gene and a second set of genes associated with the second gene, and the mapping includes (i) obtaining a second set of RNA expression levels from among the first RNA expression levels, the second set including a first RNA expression level for the second gene and RNA expression levels for genes in the second set of genes associated with the second gene; (ii) obtaining a second transformation from the RNA expression levels of one or more genes determined by the first protocol to estimate an RNA expression level for the second gene determined according to the second protocol, the second transformation being different from the first transformation; and (iii) determining a second RNA expression level for the second gene by applying the second transformation to the second set of RNA expression levels for inclusion in the second RNA expression level.

[0039] More generally, in some embodiments, the set of genes includes one or more additional genes and a further set of genes associated with the one or more additional genes, and the mapping includes (i) obtaining a set of RNA expression levels from the first RNA expression level, the set including an RNA expression level for each of at least a portion of the one or more additional genes and an RNA expression level for at least a portion of the genes of the further set of genes associated with the one or more additional genes; (ii) obtaining a respective transformation for estimating an RNA expression level for each of the one or more additional genes determined according to a second protocol; and (iii) determining a second RNA expression level for each of at least a portion of the additional genes of the subset by applying the second transformation to the first set of RNA expression levels for inclusion in the second RNA expression level.

[0040] In some embodiments, the first transformation can map the expression value of a single gene determined using a first protocol to an estimate of the RNA expression value for that single gene obtained by applying a second protocol to the same biological sample. Such a transformation can be referred to as a "one gene to one gene" or "one to one" transformation. In some embodiments, such a transformation can be a linear transformation (e.g., as shown in FIG. 2A) or any function f() that maps the expression level in the first protocol to the expression level in the second protocol, e.g., a nonlinear transformation (e.g., as shown in FIG. 2B). Examples of nonlinear transformations that can be used include transformations performed using generalized linear models, polynomial regression, random forest regression, support vector machine (SVM) regression, neural networks, gradient boosting, and / or any other suitable nonlinear regression technique. In particular, FIG. 2A shows a demonstration example of a one to one linear transformation, with a separate linear transformation used for each gene in the set of genes. For example, the RNA expression level of gene 1, 202-1, according to protocol 1, 210, is mapped using a linear transformation 204-1 to obtain a second RNA expression level of gene 1, 206-1, obtained using protocol 2, 212. In another example, the RNA expression level of gene 2, 202-2, according to protocol 1, 210, is mapped using a linear transformation 204-2 to obtain a second RNA expression level of gene 2, 206-2, obtained using protocol 2, 212. In another example, the RNA expression level of gene 3, 202-3, according to protocol 1, 210, is mapped using a linear transformation 204-3 to obtain a second RNA expression level of gene 3, 206-1, obtained using protocol 2, 212. The RNA expression level of gene N, 202-N, is mapped using a linear transformation 204-N to obtain a second RNA expression level of gene N, 206-N, obtained using protocol 2, 212. Each such linear transformation may be estimated using the paired values ​​of the expression levels for the genes.The paired values ​​of expression levels for each gene i are calculated based on whether it is a first protocol, 210 (e.g., FFPE preparation, then EC RNA-seq, "x. i " ) and a second protocol, 212 (e.g., FF preparation, followed by polyA RNA-seq, "y i A linear transformation 214 then calculates a coefficient (e.g., a ) that can be used to project the gene expression level of the gene from the first protocol to the second protocol. i and b i ) is matched between paired expression values ​​to yield

[0041] Other types of transformations (e.g., non-linear transformations) can be used as well, as shown in FIG. 2B, which illustrates that the linear transformation shown in FIG. 2A can be replaced with other types of transformations. As shown in FIG. 2B, the RNA expression levels can be transformed using any other suitable transformation f rather than the linear transformation shown in FIG. 2A. i As shown in FIG. 2B, the RNA expression level of gene 1, 214-1, according to protocol 1, 210, is mapped using function 216-1 to obtain a second RNA expression level of gene 1, 218-1, obtained using protocol 2, 212. In another example, the RNA expression level of gene 2, 214-2, according to protocol 1, 210, is mapped using function 216-2 to obtain a second RNA expression level of gene 2, 218-2, obtained using protocol 2, 212. In another example, the RNA expression level of gene 3, 214-3, according to protocol 1, 210, is mapped using function 216-3 to obtain a second RNA expression level of gene 3, 218-3, obtained using protocol 2, 212. The RNA expression level of gene N, 214-N, is mapped using function 216-N to obtain a second RNA expression level of gene N, 218-N, obtained using protocol 2, 212.

[0042] In some embodiments, the first transformation can map the RNA expression values ​​of the plurality of genes determined using a first protocol to an estimate of one RNA expression value of the plurality of genes obtained by applying a second protocol. Such a transformation can be referred to as a "many-gene-to-one-gene" or "many-to-one" transformation. A second RNA expression level 224 under the second protocol for a selected gene can be predicted from the RNA expression levels 226 for the plurality of genes obtained using the first protocol. The RNA expression level 226 includes the RNA expression level for the selected gene under the first protocol and one or more RNA expression levels for one or more genes associated with the selected gene (determined by the first protocol). In some embodiments, a separate linear transformation is used to estimate the "second protocol" RNA expression value for each gene in the set of genes. Each such linear transformation can be estimated using the paired values ​​of the RNA expression level for the gene. The estimation may have been performed in any suitable manner, including linear regression or regularized linear regression (e.g., LASSO, ridge regression, ElasticNET). Other types of transformations (e.g., non-linear transformations) may be used as well, as shown in Figure 2D, which illustrates that the linear transformation shown in Figure 2C may be replaced with other types of transformations, as aspects of the technology described herein are not limited in this respect.

[0043] In some embodiments, the multi-to-one transformation can improve the accuracy of projection compared to the single gene method using one-to-one transformation. This is because the multi-to-one transformation can utilize a combination of paired values ​​for 1) the RNA expression level of the gene of interest, and 2) the RNA expression level of the gene related to the gene of interest. In some embodiments, the gene of interest refers to the gene that is being transformed. In some embodiments, the gene related to the gene of interest is a gene that has an RNA expression level that is correlated with the expression level of the gene of interest (e.g., as determined by Pearson correlation).

[0044] Regardless of the type of transformation, in some embodiments the transformation can be estimated (using a suitable estimation technique, such as linear or non-estimation regression techniques) from the training data. As can be seen from the above, in some embodiments the training data comprises a plurality of paired values ​​of RNA expression levels for each of at least a portion of a set of genes, each pair of values ​​in the plurality of paired values ​​comprising an RNA expression level determined by applying a first protocol to a particular biological sample and another RNA expression level determined by applying a second protocol to the particular biological sample.

[0045] In some embodiments, obtaining the first set of RNA expression levels includes identifying one or more other genes associated with the first gene. In some embodiments, the identification can be performed using Pearson correlation and / or any other suitable correlation means.

[0046] In some embodiments, the first and second protocols may be different protocols for obtaining sequencing data (e.g., RNA sequencing data). The difference may be in sample preservation, preparation, sequencing, and / or any other aspect of processing the biological sample to obtain sequencing data. For example, the first protocol may include (1) preserving the biological sample by formalin fixation and paraffin embedding (FFPE) technique; and (2) performing exome capture (EC) RNA sequencing on the FFPE-preserved biological sample. As another example, the second protocol may include (1) preserving the biological sample by fresh-frozen (FF) technique; and (2) performing polyA RNA sequencing on the FF-preserved biological sample.

[0047] In some embodiments, identifying the cohort includes (1) correlating the second RNA expression level with the RNA expression level of a particular cohort of the plurality of cohorts; and (2) identifying the subject as a member of the particular cohort to which the second RNA expression level is associated.

[0048] In some embodiments, these techniques further include using the second RNA expression level to select a cancer therapeutic agent for the subject, and optionally administering the selected cancer therapeutic agent to the subject.

[0049] In some embodiments, selecting the cancer therapeutic agent includes determining the RNA expression levels of the plurality of gene groups using the second RNA expression levels, where the RNA expression levels of the plurality of gene groups include an RNA expression level of the gene group for each gene group in the set of gene groups, the set of gene groups including at least one gene group associated with cancer malignancy and at least one gene group associated with the cancer microenvironment; and selecting the cancer therapeutic agent using the determined expression levels of the gene groups.

[0050] Projection of RNA expression levels from patient-derived samples sequenced by EC RNA-seq to expression levels when the samples are prepared by polyA RNA-seq improves the fit of patient expression data with existing RNA expression data references, allowing comparison of RNA expression levels of a single sample with RNA expression levels of any other samples or cohorts of subjects, regardless of disease / non-disease status or specific disease being investigated. The ability to directly compare RNA expression data from patient samples with RNA expression data from reference datasets of large clinical studies (e.g., cancer cohort expression data, such as TCGA data), will allow researchers and physicians to better link patients with cohorts and improve the quality and accuracy of downstream analysis of patient expression data, for example, in characterizing the patient's tumor microenvironment (TME) and / or selecting cancer therapy for the patient.

[0051] FIG. 3 is a flowchart of an exemplary process 300 for mapping RNA expression levels of genes expressed in a biological sample obtained from a subject using a first protocol to the RNA expression levels that would have been determined by a second protocol if the biological sample had been processed using the second protocol instead of the first protocol, according to some embodiments of the technology described herein.

[0052] Various activities (e.g., some or all) of process 300 may be performed using any suitable computing device. For example, in some embodiments, one or more activities of exemplary process 300 may be performed in a clinical or laboratory setting. For example, one or more activities of process 300 may be performed on a computing device located within a clinical or laboratory setting. In some embodiments, the computing device may obtain expression data directly from a sequencing device located within a clinical or laboratory setting. For example, a computing device included in a sequencing device may obtain RNA expression data directly from the sequencing device. In some embodiments, the computing device may obtain RNA expression data indirectly from a sequencing device located within or outside of a clinical or laboratory setting. For example, a computing device located within a clinical or laboratory setting may obtain RNA expression data via a communications network, such as the Internet or any other suitable network, as aspects of the technology described herein are not limited to any particular communications network.

[0053] Additionally or alternatively, one or more activities of the exemplary process 300 can be performed in a setting that is separate from a clinical or laboratory setting. For example, one or more activities of the process 300 can be performed on a computing device that is located externally from the clinical or laboratory setting. In this case, the computing device can indirectly obtain RNA expression data generated using a sequencing device located inside or outside the clinical or laboratory setting. For example, the RNA expression data can be provided to the computing device via a communication network, such as the Internet or any other suitable network.

[0054] It should be understood that in some embodiments, less than all activities of process 300 as shown in Figure 3 can be performed using one or more computing devices. For example, activity 308 of using the second expression level to select a cancer therapy or select a cohort associated with the subject can be performed manually (e.g., by a clinician), automatically (e.g., by software that identifies a cancer therapy), or partially manually and partially automatically (e.g., a clinician can use information generated by the software to select a cancer therapy or cohort for the subject, e.g., using techniques described herein). In another example, activity 310 of administering a therapy to a subject can be performed manually (e.g., by a clinician).

[0055] Process 300 begins with activity 302, where first RNA expression data is obtained. The first RNA expression data may indicate (e.g., identify) that a first RNA expression level is obtained for a set of genes expressed in a biological sample obtained from a subject by a first protocol. In some embodiments, the first RNA expression level may have been determined in advance (i.e., prior to the start of process 300) by processing the biological sample using the first protocol. In other embodiments, the first protocol may be applied to the biological sample as part of activity 302.

[0056] In some embodiments, the first protocol includes: (1) preserving the biological sample using formalin fixation and paraffin embedding (FFPE); and (2) sequencing the biological sample using exome capture (EC) RNA sequencing technology to obtain a first RNA expression level. This and other examples of the first protocol are described herein, including in the sections entitled "Extracting DNA and / or RNA" and "Obtaining RNA Expression Data."

[0057] As described above, the first RNA expression data obtained in activity 302 can indicate a first RNA expression level for a set of genes. Examples of RNA expression data, sources of RNA expression data, and formats of RNA expression data are described herein, including in the section entitled "Obtaining RNA Expression Data."

[0058] The set of genes expressed in a biological sample may include any suitable number of genes present (e.g., expressed) in the biological sample. In some embodiments, the set of genes includes all genes present (e.g., expressed) in the biological sample. In some embodiments, the set of genes includes less than all genes, e.g., a subset of genes, present (e.g., expressed) in the biological sample. In some embodiments, the set of genes includes 10-25,000 genes. In some embodiments, the set of genes includes 10-1000, 500-5000, 2500-10000, 5000-15000, or 10000-25000 genes. In some embodiments, the set of genes includes 1000-2500 genes. In some embodiments, the set of genes includes or consists of the genes listed in Table 2 or Table 3. In some embodiments, the set of genes comprises or consists of at least 10%, at least 20%, at least 30%, at least 40%, at least 50%, at least 60%, at least 70%, at least 80%, at least 90%, at least 95%, at least 99%, or 100% of the genes listed in Table 2 or Table 3.

[0059] As one illustrative example, in some embodiments, the first RNA expression data may include bulk sequencing data (e.g., bulk sequencing data obtained from a single biological sample). The bulk sequencing data may include at least 1 million reads, at least 5 million reads, at least 10 million reads, at least 20 million reads, at least 50 million reads, or at least 100 million reads. In some embodiments, the sequencing data includes bulk RNA sequencing (RNA-seq) data, single-cell RNA sequencing (scRNA-seq) data, or next-generation sequencing (NGS) data. In some embodiments, the first RNA expression data includes exome capture (EC) RNA sequencing data.

[0060] Process 300 then proceeds to activity 304, where the first RNA expression levels obtained in activity 302 are mapped to second RNA expression levels for a second protocol that is different from the first protocol. For example, if the first protocol includes obtaining RNA expression levels by EC RNA-seq, the second protocol may not include obtaining EC RNA-seq expression levels, but may include, for example, obtaining polyA RNA-seq expression levels. Examples of second protocols are described herein, including in the sections entitled "Extracting DNA and / or RNA" and "Obtaining RNA Expression Data."

[0061] In activity 304, the mapping can be performed in any suitable manner as described herein. For example, in some embodiments, the mapping can include determining a projected RNA expression level for each gene in the set of genes, and for each such gene, using a respective gene-specific transformation to determine the projected gene RNA expression level. For example, if the first RNA expression level contains "N" expression levels for a set of N genes, the mapping performed in activity 304 can include projecting each of the "N" RNA expression levels using a respective transformation. As a result, "N" different transformations can be used, one for each of the N genes. Each such transformation can be a one-to-one transformation (e.g., see Figures 2A and 2B) or a many-to-one transformation (e.g., see Figures 2C and 2D). In some embodiments, each such transformation can be linear. In some embodiments, each such transformation is independently a linear or nonlinear transformation (e.g., a first linear transformation and a second nonlinear transformation). In some embodiments, each such transformation may be estimated (i.e., the parameters of the transformation were determined) from the training data (including the paired values ​​described herein) using any suitable estimation technique (e.g., linear regression or regularized linear regression, examples of which are provided herein). A "projected" RNA expression level refers to the estimated RNA expression level for a gene in the set of genes expressed in a biological sample that would have been determined by a second protocol if the biological sample were processed using the second protocol instead of the first protocol. Aspects of the mapping performed in activity 304 are described herein, including with reference to FIG. 4.

[0062] In some embodiments, process 300 may be completed after activity 304 is completed. In other embodiments, process 300 may continue and perform one or more of optional activities 306, 308, and 310. For example, only activity 306 may be performed, or only activity 308 may be performed, or both activities 306 and 308 may be performed, or both activities 308 and 310 may be performed, or all three activities 306, 308, and 310 may be performed.

[0063] In activity 306, the second RNA expression level obtained as a result of the mapping performed in activity 304 is used to identify a cohort associated with the subject from which the biological sample was obtained. Embodiments of methods for identifying a cohort using the second RNA expression level are described herein, including in the section entitled "Post-Mapping Processing."

[0064] In activity 308, the second RNA expression level can be used to select a cancer therapy, and in activity 310, the selected therapy can be administered to the subject. Embodiments of how activities 308 and 310 can be performed are described herein, including in the sections titled "Post-Mapping Processing" and "Anti-Cancer Therapy."

[0065] 4 is a flow chart depicting an exemplary process 400 for mapping RNA expression levels obtained using a first protocol to RNA expression levels obtained using a second, different protocol, according to some embodiments of the technology described herein. Process 400 can be used to perform activity 304 described with reference to process 300. Process 400 can be performed using any computing device, as aspects of the technology described herein are not limited in this respect.

[0066] Process 400 begins at activity 402, where a particular gene is selected from a set of genes. Examples of genes and sets of genes are provided herein.

[0067] Process 400 then proceeds to activity 404, where a set of RNA expression levels is obtained for the selected gene. The RNA expression levels may have been determined by applying a first protocol (e.g., EC RNA-seq) to a biological sample obtained from the subject. As shown in FIG. 4, the set of RNA expression levels may include a single RNA expression level that may be obtained at activity 404a, which may be an RNA expression level for the gene selected at activity 402. Optionally, the set of RNA expression levels may include one or more additional RNA expression levels that may be obtained at activity 404b for one or more other genes related to the gene selected at activity 402.

[0068] In some embodiments, the one or more other genes may be any suitable number of genes. In some embodiments, the plurality of genes includes 1-10, 5-20, 10-50, 25-100, 50-200, 125-500, 250-1000, or any other range within these ranges or more than 1000 genes. In some embodiments, the one or more RNA expression levels of the one or more other genes include 1-10, 5-20, 10-50, 25-100, 50-200, 125-500, 250-1000, or any other range within these ranges or more than 1000 genes.

[0069] Genes that are "associated with" a selected gene are genes that have RNA expression levels that are correlated with the RNA expression levels of the selected gene. The correlation of RNA expression levels can be measured by any suitable method known in the art. Examples of techniques that can be used to identify the association between RNA expression levels include, but are not limited to, Pearson correlation. Thus, in some embodiments, for each particular gene, genes that are "associated with" the particular gene can be identified by Pearson correlation.

[0070] Process 400 then proceeds to activity 406, where a transformation for the selected genes is obtained. In some embodiments, the transformation is pre-determined (e.g., determined before the start of process 400). In some embodiments, the transformation may be a linear transformation, while in other embodiments, a non-linear transformation may be used.

[0071] In some embodiments, the transformation may be pre-determined from the training data by using any suitable linear (or non-linear) regression technique. For example, linear regression (e.g., ordinary least squares (OLS)) or regularized linear regression (LASSO, ridge regression, ElasticNet or ElasticNetCV regression) may be used. ElasticNet or ElasticNetCV regression is described by Zou and Hastie, 2005, "Regularization and variable selection via the elastic net." Journal of the Royal Statistical Society. Series B, Statistical methodology 67(2):301-320, which is incorporated herein by reference in its entirety.

[0072] In some embodiments, the training data includes paired values ​​of RNA expression levels for selected genes of the set of RNA expression data. Each paired value of the RNA expression level may include an RNA expression level determined by applying a first protocol to a particular biological sample (e.g., an RNA expression level of protocol 1) and another RNA expression level determined by applying a second protocol to the particular biological sample (e.g., an RNA expression level of protocol 2). The training data (for each gene) may include any suitable number of training values ​​(e.g., at least 5, 10, 100, 1000, 5000, 10,000, 5-1000, 100-10,000 paired values, or any other suitable range within these ranges). The training data may include paired values ​​of RNA expression levels for selected genes in a single sample (e.g., all paired values ​​of RNA expression levels are obtained from a single biological sample) or RNA expression levels for selected genes in multiple biological samples (e.g., paired RNA expression levels are obtained from multiple biological samples, such as 1, 2, 5, 10, 100, 500, 1000, 5000, or 10000 samples).

[0073] Process 400 then proceeds to activity 408, where the transformation selected at activity 406 is applied to the set of RNA expression levels obtained at activity 404 to obtain projected "Protocol 2" RNA expression levels for the selected genes. The projected "Protocol 2" RNA expression levels for the selected genes indicate the RNA expression levels of the selected genes in the biological sample if the biological sample were processed according to the second protocol rather than the first protocol.

[0074] Process 400 then proceeds to activity 410 to determine whether to repeat activities 404-408. If the RNA expression levels of other genes in the biological sample are not to be mapped, process 400 terminates at activity 410. If the RNA expression levels of one or more additional genes are to be mapped, process 400 returns to activity 402 to select another gene for mapping and repeat activities 404-410.

[0075] The number of genes in a biological sample having RNA expression levels that are mapped from Protocol 1 to Protocol 2 may vary. In some embodiments, all of the genes in a biological sample are mapped using process 400. In some embodiments, less than all of the genes in a biological sample (e.g., a subset of the genes) are mapped using process 410. The subset may have 10-25,000 genes, 10-1000, 500-5000, 2500-10000, 5000-15000, or 10000-25000 genes. In some embodiments, the subset of genes includes 1000-2500 genes. In some embodiments, the subset includes or consists of the genes listed in Table 2 or Table 3.

[0076] Biological samples Aspects of the present disclosure relate to methods for mapping RNA expression levels of genes expressed in a biological sample obtained from a subject using a first protocol to the RNA expression levels that would be determined by a second protocol if the biological sample was processed using the second protocol instead of the first protocol.

[0077] In some embodiments, the subject is a mammal (e.g., a human, mouse, cat, dog, horse, hamster, cow, pig, or other domestic animal). In some embodiments, the subject is a human. In some embodiments, the subject is an adult (e.g., 18 years of age or older). In some embodiments, the subject is a child (e.g., under 18 years of age). In some embodiments, the human subject has or has been diagnosed with at least one form of cancer. In some embodiments, the cancer the subject suffers from is carcinoma, sarcoma, myeloma, leukemia, lymphoma, or a mixed type cancer comprising two or more of carcinoma, sarcoma, myeloma, leukemia, and lymphoma. Carcinoma refers to a malignant neoplasm of epithelial origin or a cancer of the lining or lining of the body. Sarcoma refers to a cancer originating from supporting and connective tissues such as bone, tendon, cartilage, muscle, and fat. Myeloma is a cancer originating from the plasma cells of the bone marrow. Leukemia ("liquid cancer" or "blood cancer") is a cancer of the bone marrow (site of blood cell production). Lymphomas arise in the glands or nodes, vascular networks, nodes, and organs of the lymphatic system (specifically the spleen, tonsils, and thymus) that produce white blood cells, or lymphocytes, that purify body fluids and fight infection. Non-limiting examples of mixed cancers include adenosquamous carcinoma, mixed mesodermal tumor, carcinosarcoma, and teratocarcinoma. In some embodiments, the subject has a tumor. The tumor may be benign or malignant. In some embodiments, the cancer is any one of the following: skin cancer, lung cancer, breast cancer, prostate cancer, colon cancer, rectal cancer, cervical cancer, and uterine cancer. In some embodiments, the subject is at risk of developing cancer because, for example, the subject has one or more genetic risk factors or has been or is exposed to one or more carcinogens (e.g., smoking or chewing tobacco).

[0078] The present disclosure is based, in part, on projecting the RNA expression level of a gene in a biological sample prepared according to a first protocol to the RNA expression level of a gene in a biological sample when the sample is prepared by a second protocol (e.g., a protocol different from the first protocol). As used herein, the term "protocol" refers to one or more techniques used to obtain, isolate, preserve, or process a biological sample obtained from a subject. Examples of techniques for obtaining tissue from a subject include, but are not limited to, liquid (e.g., blood, CSF, lymph nodes, etc.) collection, tissue biopsy, cell exfoliation, urine sample collection, fecal sample collection, saliva collection, etc. Examples of methods for preserving biological samples include, but are not limited to, fresh frozen preservation techniques and tissue fixation techniques (e.g., alcohol fixation, formalin fixation, paraffin embedding, optimal cutting temperature (OCT) preservation, RNAlater® preservation, etc.). Examples of processing techniques include, but are not limited to, nucleic acid extraction, nucleic acid purification, and nucleic acid sequencing.

[0079] In some embodiments, the RNA expression data is obtained from biological samples prepared by a protocol including formalin fixation and paraffin embedding (FFPE). Examples of FFPE techniques include, but are not limited to, laser capture microdissection (LCM), microtome sectioning, and FFPE core isolation. Methods for FFPE preservation of tissues are well known and are described, for example, in Amini et al., BMC Molecular Biology volume 18, Article number: 22 (2017). Typically, FFPE protocols include the following steps: tissue coring, tissue fixation, paraffin embedding, mounting, and storage. FFPE-preserved samples can be stored at room temperature or below room temperature, for example, at 4°C.

[0080] In some embodiments, the protocol comprising FFPE preservation further comprises nucleic acid extraction and / or nucleic acid purification. Examples of nucleic acid extraction and purification techniques are described herein in the section entitled "DNA and / or RNA Extraction". In some embodiments, the protocol comprising FFPE preservation further comprises nucleic acid sequencing. In some embodiments, the nucleic acid sequencing is exome capture (EC) RNA sequencing (RNA-seq). Methods of sequencing, including EC RNA-seq, are described herein, including in the section entitled "Obtaining Gene Expression Data".

[0081] In some embodiments, RNA expression data is obtained from biological samples prepared by a protocol that includes fresh frozen preservation technology.The method for preserving fresh frozen tissue generally includes the following steps: tissue collection, quick freezing by immersion in liquid nitrogen, and storage at -80°C, as described, for example, by Mager et al., Standard operating procedure for the collection of fresh frozen tissue samples.Eur J Cancer 2007, 43(5):828-834.

[0082] In some embodiments, the protocol comprising FF storage further comprises nucleic acid extraction and / or nucleic acid purification. Examples of nucleic acid extraction and purification techniques are described herein in the section entitled "Extraction of DNA and / or RNA". In some embodiments, the protocol comprising FF storage further comprises nucleic acid sequencing. In some embodiments, the nucleic acid sequencing is polyA RNA-seq. Methods of sequencing, including polyA RNA-seq, are described herein, including in the section entitled "Obtaining Gene Expression Data".

[0083] The biological sample may be from any source within the subject's body, including, but not limited to, any fluid such as blood (e.g., whole blood, serum, or plasma), lymph nodes, stomach, small intestine. Other sources within the subject's body may be from saliva, tears, synovial fluid, cerebrospinal fluid, pleural fluid, pericardial fluid, peritoneal fluid, and / or urine, hair, skin (including portions of the epidermis, dermis, and / or subcutaneous tissue), oropharynx, laryngopharynx, esophagus, bronchi, salivary glands, tongue, oral cavity, nasal cavity, vaginal cavity, anal cavity, bone, bone marrow, brain, thymus, spleen, appendix, colon, rectum, anus, liver, bile duct, pancreas, kidney, ureter, bladder, urethra, uterus, vagina, vulva, ovaries, cervix, scrotum, penis, prostate, testes, seminal vesicles, and / or any type of tissue (e.g., muscle tissue, epithelial tissue, connective tissue, or nerve tissue). A biological sample may be any type of sample, including, for example, a sample of a body fluid, one or more cells, or one or more tissue or organ slices.

[0084] In some embodiments, a tissue sample can be obtained from a subject using a surgical procedure, a bone marrow biopsy, a punch biopsy, an endoscopic biopsy, or a needle biopsy (e.g., fine needle aspiration, core needle biopsy, vacuum assisted biopsy, or image-guided biopsy).

[0085] A lymph node or blood sample, in some embodiments, refers to a sample that includes cells, e.g., cells derived from a blood sample or a lymph node sample. In some embodiments, the sample includes non-cancerous cells. In some embodiments, the sample includes pre-cancerous cells. In some embodiments, the sample includes cancerous cells. In some embodiments, the sample includes blood cells. In some embodiments, the sample includes lymph node cells. In some embodiments, the sample includes lymph node cells and blood cells.

[0086] The blood sample may be a whole blood sample or a fractionated blood sample. In some embodiments, the blood sample comprises whole blood. In some embodiments, the blood sample comprises fractionated blood. In some embodiments, the blood sample comprises a buffy coat. In some embodiments, the blood sample comprises serum. In some embodiments, the blood sample comprises plasma. In some embodiments, the blood sample comprises a thrombus.

[0087] In some embodiments, a sample of blood is collected to obtain cell-free nucleic acid (eg, cell-free DNA) in the blood.

[0088] In some embodiments, the sample may be derived from a cancerous tissue or organ or tissues or tissues or organs suspected of having one or more cancerous cells. In some embodiments, the sample may be derived from a healthy (e.g., non-cancerous) tissue or organ. In some embodiments, a sample from a subject (e.g., a biopsy from a subject) may include both healthy cells and / or tissues and cancerous cells and / or tissues. In certain embodiments, one sample will be taken from a subject for analysis. In some embodiments, more than one (e.g., 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20 or more) samples can be taken from a subject for analysis. In some embodiments, one sample from a subject will be analyzed. In certain embodiments, more than one (e.g., 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20 or more) sample can be analyzed. When analyzing more than one sample from a subject, the samples can be produced simultaneously (e.g., more than one sample can be taken in the same procedure) or the samples can be taken at different times (e.g., between different procedures, including 1, 2, 3, 4, 5, 6, 7, 8, 9, 10 days after the first procedure; 1, 2, 3, 4, 5, 6, 7, 8, 9, 10 weeks after; 1, 2, 3, 4, 5, 6, 7, 8, 9, 10 months after, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10 years after, or 10, 20, 30, 40, 50, 60, 70, 80, 90, 100 years after the procedure). The second or subsequent sample may be taken or obtained from the same area (e.g., the same tumor or tissue area) or a different area (e.g., including a different tumor). The second or subsequent sample may be taken or obtained from the subject after one or more treatments and may be taken from the same area or a different area. As a non-limiting example, the second or subsequent sample may be useful to determine whether the cancer in each sample has different characteristics (e.g., in the case of samples taken from two physically separate tumors in a patient) or whether the cancer has responded to one or more treatments (e.g., in the case of two or more samples from the same tumor before and after treatment).

[0089] Any biological sample described herein can be obtained from a subject using any known technique.See, for example, the following publications on biological sample collection, processing, and storage: Biospecimens and biorepositories: from afterthought to science by Vaught et al. (Cancer Epidemiol Biomarkers Prev. 2012 Feb; 21(2): 253-5), and Biological sample collection, processing, storage and information management by Vaught and Henderson (IARC Sci Publ. 2011; (163): 23-42), each of which is incorporated herein by reference in its entirety.

[0090] Any biological sample derived from a subject described herein can be stored using any method that maintains the stability of the biological sample. In some embodiments, maintaining the stability of the biological sample means that the components of the biological sample (e.g., DNA, RNA, protein, or tissue structure or morphology) are not degraded until they are measured, so that when measured, the measurement represents the state of the sample at the time it is obtained from the subject. In some embodiments, the biological sample is stored in a composition that can permeate it and protect the components of the biological sample (e.g., DNA, RNA, protein, or tissue structure or morphology) from degradation. As used herein, degradation is the conversion of a component from one form to another, such that the first form is no longer detected at the same level as before degradation.

[0091] In some embodiments, the biological sample is stored using cryopreservation. Non-limiting examples of cryopreservation include, but are not limited to, step-down freezing, rapid freezing, direct plunge freezing, quick freezing, slow freezing using a programmable freezer, and vitrification. In some embodiments, the biological sample is stored using freeze-drying. In some embodiments, the biological sample is frozen (e.g., by quick freezing) after collection of the biological sample from the subject, after being placed in a container that already contains a preservative (e.g., RNALater for preserving RNA). In some embodiments, such storage in a frozen state is performed immediately after collection of the biological sample. In some embodiments, the biological sample may be kept at room temperature or 4°C for some time (e.g., up to 1 hour, up to 8h, or up to 1 day, or up to several days) in a preservative or in a buffer that does not contain a preservative before freezing.

[0092] Non-limiting examples of preservatives include formalin solution, formaldehyde solution, RNALater or other equivalent solution, TriZol or other equivalent solution, DNA / RNA Shield or equivalent solution, EDTA (e.g., Buffer AE (10 mM Tris Cl; 0.5 mM EDTA, pH 9.0)) and other coagulants, and citrate dextrose (e.g., for blood specimens).

[0093] In some embodiments, specialized containers can be used to collect and / or store biological samples. For example, a vacutainer can be used to store blood. In some embodiments, the vacutainer may contain a preservative (e.g., a clotting agent, or an anticoagulant). In some embodiments, the container in which the biological sample is stored may be contained within a secondary container for better storage or to avoid contamination.

[0094] DNA and / or RNA extraction In some embodiments of any one of the methods described herein, RNA is extracted from a biological sample to prevent it from being degraded and / or to prevent enzyme inhibition during downstream processing, such as the preparation of DNA (i.e., a cDNA library derived from the RNA). In some embodiments, the term "extraction" in the context of obtaining RNA from a biological sample is used interchangeably with the term "isolation."

[0095] The methods described herein include extraction of RNA from a biological sample (e.g., a tumor sample or a blood sample). As described above, a biological sample may include more than one sample from one or more tissues (e.g., one or more different tumors). In some embodiments, RNA is extracted from the combined samples. In some embodiments, RNA is extracted from multiple biological samples from a subject and then combined prior to further processing (e.g., storage, or DNA library preparation). In some embodiments, more than one sample of extracted RNA is combined with each other after retrieval from storage. In some embodiments, at least tumor is extracted from one or more tumor tissues. In some embodiments, at least tumor RNA is extracted from one or more tumor tissues. In some embodiments, at least normal RNA is extracted from one or more normal tissues. In some embodiments, RNA is extracted from a normal sample that serves as a control.

[0096] Methods for extracting RNA from biological samples are known, and reagents and kits for doing so are commercially available. Gomez-Acata et al. (Methods for extracting 'omes from microbialites, J Microbiol Methods. 2019 Mar 12; 160: pp. 1-10) describe extraction methods applied for RNA extraction from microbialites, and describe their advantages and disadvantages, which are incorporated herein by reference in their entirety. The methods described in Gomez-Acata et al. are generally applicable to RNA extracted from tissues. Dowhan (Curr. Protoc. Essential Lab. Tech. 6: 5.2.1-5.2.21) describes the purification and concentration of RNA from aqueous solutions, which are also incorporated herein by reference in their entirety. In some embodiments, RNA is extracted from the biological sample using a kit suitable for RNA-seq, for example by the method described in Cortes-Esteve et al., PLoS One. 2017;12(1):e0170632.

[0097] In some embodiments, extracting RNA comprises lysing the cells of biological samples and isolating RNA from other cellular components.Examples of methods for lysing cells include, but are not limited to, mechanical lysis, liquid homogenization, sonication, freeze-thaw, chemical lysis, alkaline lysis, and manual grinding.

[0098] Methods for extracting RNA include, but are not limited to, solution phase extraction and solid phase extraction. In some embodiments, the solution phase extraction comprises an organic extraction, such as phenol-chloroform extraction. In some embodiments, the solution phase extraction comprises a high salt extraction, such as guanidinium thiocyanate (GuTC) or guanidinium chloride (GuCl) extraction. In some embodiments, the solution phase extraction comprises ethanol precipitation. In some embodiments, the solution phase extraction comprises isopropanol precipitation. In some embodiments, the solution phase extraction comprises ethidium bromide (EtBr)-cesium chloride (CsCl) gradient centrifugation. In some embodiments, the extraction of DNA and / or RNA comprises a non-ionic detergent extraction, such as cetyltrimethylammonium bromide (CTAB) extraction.

[0099] In some embodiments, the extraction of RNA comprises solid-phase extraction.Any solid phase that binds RNA can be used to extract RNA in the methods and systems described herein.The examples of solid phase that binds RNA include, but are not limited to, silica matrix, ion exchange matrix, glass particles, magnetizable cellulose beads, polyamide matrix, and nitrocellulose membrane.

[0100] In some embodiments, the solid phase extraction method comprises a spin column-based extraction method. In some embodiments, the solid phase extraction method comprises a bead-based extraction method. In some embodiments, the solid phase extraction method comprises a cation exchange resin, such as a styrene divinyl benzene copolymer resin.

[0101] The systems and methods described herein encompass extracting RNA from a single biological sample or multiple biological samples. In some embodiments, extracting RNA comprises extracting RNA from a single sample. In some embodiments, extracting RNA comprises extracting RNA from multiple samples. In some embodiments, extracting RNA comprises extracting RNA from a first sample and a second sample. In some embodiments, extracting RNA comprises extracting RNA from 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 samples.

[0102] The RNA extracted from a biological sample can be combined with the RNA extracted from another biological sample. This can be achieved by combining one or more biological samples and extracting nucleic acid, or by combining the nucleic acid extracted from one or more biological samples. In some embodiments, the first biological sample is combined with the second biological sample to form a combined sample, and RNA is extracted from the combined sample. In some embodiments, the RNA extracted from the first biological sample can be combined with the DNA and / or RNA extracted from the second biological sample.

[0103] The systems and methods described herein encompass extracting any type of RNA from a biological sample. In some embodiments, extracting RNA comprises extracting messenger RNA (mRNA). In some embodiments, extracting RNA comprises extracting precursor mRNA (pre-mRNA). In some embodiments, extracting RNA comprises extracting ribosomal RNA (rRNA). In some embodiments, extracting RNA comprises extracting transfer RNA (tRNA).

[0104] In some embodiments, a single kit is used to purify DNA and RNA from the same sample. A non-limiting example of a kit for doing so is the Qiagen AllPrep DNA / RNA kit. In some embodiments, robotics is used to perform DNA and / or RNA extraction.

[0105] In some embodiments, the extracted RNA is checked for RNA quality and / or quantity before further processing for RNA sequencing or whole exome sequencing (WES). In some embodiments, the extracted RNA sample is at least 1000-6000 ng in total mass. In some embodiments, the extracted RNA sample is at least 100-60000 ng in total mass (e.g., 100-60000 ng, 500-30000 ng, 800-20000 ng, 1000-15000 ng, 1000-10000 ng, 1000-8000 ng, 1000-6000 ng, 10000-20000 ng, 20000-60000 ng). In some embodiments, the acceptable total RNA amount for further sequencing is at least 100-1,000 ng (e.g., 100-1,000 ng, 500-1,000 ng, or 300-900 ng). In some embodiments, the target total RNA amount for further sequencing is more than 200-1,000 ng (e.g., 200-1,000 ng, 500-1,000 ng, or 300-1,000 ng). In some embodiments, the purity of the sample of extracted RNA is such that it corresponds to a ratio of absorbance at 260 nm to absorbance at 280 nm of at least 1 (e.g., at least 1, at least 1.2, at least 1.4, at least 1.6, at least 1.8, or at least 2). In some embodiments, the purity of the sample of extracted RNA is such that it corresponds to a ratio of absorbance at 260 nm to absorbance at 280 nm of at least 2. The ratio of absorbance at 260 nm and 280 nm has been used to assess the purity of DNA and RNA. A ratio of about 1.8 is generally accepted as "pure" for DNA; a ratio of about 2.0 is generally accepted as "pure" for RNA. If the ratio is significantly lower in either case, it may indicate the presence of proteins, phenols, or other contaminants that absorb strongly at or near 280 nm. Absorbance can be measured using a spectrophotometer.

[0106] In some embodiments, the purity or integrity of the extracted RNA is such that it corresponds to an RNA integrity number (RIN) of at least 4 (e.g., at least 4, at least 5, at least 6, at least 7, at least 8, or at least 9). In some embodiments, the purity of the extracted RNA is such that it corresponds to an RNA integrity number (RIN) of at least 7. RIN has proven robust and reproducible in studies comparing it to other RNA integrity calculation algorithms and solidifying its position as the preferred method for determining the quality of the RNA to be analyzed (Imbeaud et al., Towards standardization of RNA quality assessment using user-independent classifiers of microcapillary electrophoresis traces; Nucleic Acids Research. 33(6):e56).

[0107] In some embodiments, the extracted RNA sample has a target concentration of at least 2ng / μl (e.g., 2ng / μl, 4ng / μl, 6ng / μl). In some embodiments, the extracted RNA sample has an acceptable concentration of at least 4ng / μl (e.g., 4ng / μl, 6ng / μl, 10ng / μl). In some embodiments, the concentration of the extracted DNA is performed by a fluorometer (e.g., Qubit fluorometer available from ThermoFisher Scientific, www.thermofisher.com), for example, to quantify the RNA.

[0108] In some embodiments, the extracted RNA sample has a target concentration of at least 4ng / μl (e.g., 4ng / μl, 6ng / μl, 8ng / μl). In some embodiments, the extracted RNA sample has an acceptable concentration of at least 1.5ng / μl (e.g., 1.5ng / μl, 3.5ng / μl, 5.5ng / μl). In some embodiments, the extracted RNA concentration is performed by Tapestation. In some embodiments, the acceptable RNA Integrity Number (RIN) is at least 5 (e.g., 5, 6, 7). In some embodiments, the target RNA Integrity Number (RIN) is at least 8 (e.g., 8, 9, 10). In some embodiments, the RIN is performed by Tapestation.

[0109] In some embodiments, the target purity of a sample of extracted RNA is such that it corresponds to a range of ratios of absorbance at 260 nm to absorbance at 280 nm of at least 1.8 to 2 (e.g., at least 1.8 to 2, at least 1.8 to 1.9). In some embodiments, the purity of a sample of extracted RNA is such that it corresponds to a ratio of absorbance at 260 nm to absorbance at 280 nm of at least 1.8. In some embodiments, an acceptable purity of a sample of extracted RNA is such that it corresponds to a ratio of absorbance at 260 nm to absorbance at 280 nm of at least 1.5 (e.g., at least 1.5, at least 1.7, at least 2). In some embodiments, the target purity of a sample of extracted RNA is such that it corresponds to a range of ratios of absorbance at 260 nm to absorbance at 230 nm of at least 2 to 2.2 (e.g., at least 2 to 2.2, at least 2 to 2.1). In some embodiments, an acceptable purity of a sample of extracted RNA is such that it corresponds to a ratio of absorbance at 260 nm to absorbance at 230 nm of at least 1.5 (e.g., at least 1.5, at least 1.7, at least 2). In some embodiments, the purity of a sample of extracted RNA described herein is analyzed by a spectrophotometer, for example, a small volume full spectrum, UV-visible spectrophotometer (e.g., a Nanodrop spectrophotometer available from ThermoFisher Scientific).

[0110] In some embodiments, the purity of the sample of extracted RNA described herein can be analyzed by any other suitable technique or means.In some embodiments, if the sample of extracted RNA or DNA does not meet the above-mentioned specific quantitative or purity criteria, it is not further processed.In some embodiments, if the sample of extracted RNA does not meet the specific quantitative or purity criteria, it is combined with another sample.

[0111] Acquiring RNA expression data Aspects of the present disclosure relate to methods of determining RNA expression levels of genes in a subject using sequencing or RNA expression data obtained from a biological sample derived from a subject.

[0112] Sequencing data can be obtained from biological samples using any suitable sequencing technology and / or device. In some embodiments, the sequencing device used to sequence biological samples can be selected from any suitable known sequencing device, including but not limited to Illumina™, SOLid™, Ion Torrent™, PacBio™, nanopore-based sequencing device, Sanger sequencing device, or 454™ sequencing device. In some embodiments, the sequencing device or technology used to sequence biological samples is an Illumina sequencing (e.g., TrueSeq™, NovaSeq™, NextSeq™, HiSeq™, MiSeq™, or MiniSeq™) device or technology. In some embodiments, the sequencing device or technology used to sequence the biological sample is an Agilent sequencing device or technology (e.g., SureSelect™) or a NimbleGen sequencing device or technology, e.g., as described by Sulonen et al., Comparison of solution-based exome capture methods for next generation sequencing, Genome Biol 12, R94(2011).doi.org / 10.1186 / gb-2011-12-9-r94.

[0113] In some embodiments, the term "RNA sequencing" can be used interchangeably with "RNA seq", "RNA-seq", or variations thereof as known to refer to any technology, means, or platform that interrogates the transcriptome. It is understood that when "RNA sequencing", "RNA seq", "RNA-seq", or variations thereof are referred to in this disclosure, it does not refer to a specific technology or means associated with a specific platform or company, unless otherwise indicated as a limiting example only to illustrate the process or system described herein. In some embodiments, RNA sequencing can be performed by using any suitable sequencing platform and / or sequencing method. Non-limiting examples of high-efficiency sequencing platforms include mRNA-seq, total RNA-seq, targeted RNA-seq, single-cell RNA-seq, RNA exome capture platform, or small RNA-seq (e.g., Illumina, www.illumina.com), SMRT (single molecule, real-time) sequencing (e.g., Pacific Biosciences), and RNA sequencing (e.g., ThermoFisher).

[0114] As mentioned above, RNA sequencing can be targeted or non-targeted.Targeted methods include using sequence-specific probes or oligonucleotides to sequence one or more specific regions of transcriptome.In some embodiments, targeted RNA sequencing includes methods such as mRNA enrichment (e.g., by polyA enrichment or rRNA depletion).

[0115] In some embodiments, RNA sequencing is whole transcriptome sequencing. Whole transcriptome sequencing includes measuring the complete complement of transcripts in a sample. In some embodiments, whole transcriptome sequencing is used to determine the overall expression level of each transcript (both coding and non-coding) and identify exons, introns and / or their junctions.

[0116] In some embodiments, RNA is sequenced directly without preparing cDNA from a sample of RNA. In some embodiments, direct RNA sequencing comprises single molecule RNA sequencing (DRS™).

[0117] In some embodiments, the RNA sequencing is mRNA sequencing. In some embodiments, the mRNA sequencing is sequencing of only coding transcripts with the goal of removing non-coding regions. In some embodiments, the mRNA sequencing is independent of polyA enrichment. In some embodiments, the mRNA sequencing is dependent on polyA enrichment.

[0118] In some embodiments, RNA is extracted from a biological sample, mRNA is enriched from the extracted RNA, and a cDNA library is constructed from the enriched mRNA. In some embodiments, a single piece (e.g., molecule) of cDNA from a cDNA library is bound to a solid matrix. In some embodiments, a single piece (e.g., molecule) of cDNA from a cDNA library is bound to a solid matrix by limiting dilution. In some embodiments, the matrix-bound cDNA pieces (e.g., molecules) are then sequenced (e.g., using Pacbio or Pacifbio technology). In some embodiments, the matrix-bound cDNA pieces (e.g., molecules) are amplified and sequenced (e.g., using SOLiD, 454 Pyrosequencing, Ion Torrent, or special emulsion PCR (emPCR) in a connector based bridge reaction (Illumina) platform).

[0119] In some embodiments, the cDNA transcripts can be sequenced simultaneously by measuring the incorporation of fluorescent nucleotides (e.g., Illumina), fluorescent short linkers (e.g., SOLiD), by the release of by-products derived from the incorporation of normal nucleotides (454), by measuring fluorescence emission, or by measuring pH changes (e.g., Ion Torrent). In some embodiments, the cDNA transcripts can be sequenced using any known sequencing platform. Jazayeri et al. (RNA-seq: a glance at technologies and methodologies; Acta biol. Colomb. vol. 20, no. 2 Bogota May / Aug. 2015) provides a comparison of different RNA-seq platforms, which is incorporated herein by reference in its entirety, including the RNA-seq techniques listed in Tables 3 and 4. Mestan et al. (Genomic sequencing in clinical trials; Journal of Translational Medicine 2011, 9:222) provides a similar analysis for sequencing in clinical trials.

[0120] In some embodiments, RNA sequencing is stranded or strand-specific. cDNA synthesis from RNA results in loss of strandedness. In some embodiments, strandedness is maintained by chemically labeling one or both of the RNA strand and the cDNA strand formed by reverse transcription or antisense transcription, or by using adapter-based techniques to distinguish the original RNA strand from the complementary DNA strand, as described above.

[0121] In some embodiments, non-stranded RNA sequencing is performed. In some embodiments, stranded RNA-seq is not preferred for clinical samples. In some embodiments, non-stranded RNA-seq is used to compare the data obtained from biological samples with the RNA sequencing data in established data sets (e.g., The Cancer Genome Atlas (TCGA) and International Cancer Genome Consortium (ICGC)).

[0122] In some embodiments, RNA sequencing produces paired-end reads. Paired-end reads are reads of the same nucleic acid fragment, starting from either end of the fragment. In some embodiments, RNA sequencing is performed using paired-end reads of at least 2x25 (2x25, 2x50, 2x75, 2x100, 2x125, 2x150, 2x175, 2x200, 2x225, 2x250, 2x275, 2x300, 2x325, or 2x350) paired-end reads. In some embodiments, RNA sequencing is performed using paired-end reads of at least 2x75 paired-end reads. RNA sequencing using 2x75 paired-end reads means that, on average, each paired-end read reads 75 base pairs. In some embodiments, RNA sequencing is performed using a total of at least 20 million paired-end reads (e.g., at least 20 million, at least 30 million, at least 40 million, at least 50 million, at least 60 million, at least 70 million, at least 80 million, at least 90 million, at least 100 million, at least 120 million, at least 140 million, at least 150 million, at least 160 million, at least 180 million, at least 200 million, at least 250 million, at least 300 million, at least 350 million, or at least 400 million). In some embodiments, RNA sequencing is performed using a total of at least 50 million paired-end reads. In some embodiments, RNA sequencing is performed using a total of at least 100 million paired-end reads.

[0123] In some embodiments, quality control is performed for RNA sequencing. In some embodiments, cluster density or cluster PF% is a parameter for determining the quality of a sample run. In some embodiments, the target range of cluster density or cluster PF% is at least 170-220 (e.g., 170-220, 190-220, 210-220). In some embodiments, the acceptable range of cluster density or cluster PF% is at least 280 (e.g., 280, 300, 450).

[0124] In some embodiments, %≧Q30 is a parameter for determining the quality of a sample run. In some embodiments, the target %≧Q30 is at least 85% (e.g., 85%, 90%, 95%). In some embodiments, the acceptable %≧Q30 is at least 75% (e.g., 75%, 85%, 95%).

[0125] In some embodiments, the error rate % is a parameter for determining the quality of a sample run. In some embodiments, the target error rate % is less than 0.7% (e.g., 0.6%, 0.5%, 0.4%). In some embodiments, the acceptable error rate % is less than 1% (e.g., 0.9%, 0.8%, 0.7%).

[0126] After obtaining sequencing data, it is processed to obtain RNA expression data.RNA expression data can be obtained by using any known method, including but not limited to whole transcriptome sequencing, whole exome sequencing, total RNA sequencing, mRNA sequencing, targeted RNA sequencing, RNA exome capture sequencing, next generation sequencing, and / or deep RNA sequencing.In some embodiments, RNA expression data can be obtained by using microarray assay.

[0127] In some embodiments, the sequencing data is processed to produce RNA expression data. In some embodiments, the RNA sequence data is processed by one or more bioinformatics methods or software tools, such as RNA sequence quantification tools (e.g., Kallisto) and genome annotation tools (e.g., Gencode v23) to produce expression data. The Kallisto software is described in Nicolas L Bray, Harold Pimentel, Pall Melsted and Lior Pachter, Near-optimal probabilistic RNA-seq quantification, Nature Biotechnology 34, pp. 525-527 (2016), doi:10.1038 / nbt.3519, which is incorporated herein by reference in its entirety.

[0128] In some embodiments, the microarray expression data is processed using a bioinformatics R package such as "affy" or "limma" to produce expression data. The "affy" software is described in Bioinformatics. 2004 Feb 12;20(3):307-15, doi:10.1093 / bioinformatics / btg405. "affy-analysis of Affymetrix GeneChip data at the probe level", Laurent Gautier 1, Leslie Cope, Benjamin M Bolstad, Rafael A Irizarry PMID:14960456 DOI:10.1093 / bioinformatics / btg405, which is incorporated herein by reference in its entirety. The "limma" software is described in Ritchie ME, Phipson B, Wu D, Hu Y, Law CW, Shi W, Smyth GK, "limma powers differential expression analyses for RNA-sequencing and microarray studies." Nucleic Acids Res. 2015 Apr 20;43(7):e47.20. https: / / doi.org / 10.1093 / nar / gkv007 PMID:25605792, PMCID:PMC4402510, which is incorporated by reference in its entirety.

[0129] In some embodiments, the sequencing data and / or RNA expression data comprises more than 5 kilobases (kb). In some embodiments, the size of the obtained RNA data is at least 10 kb. In some embodiments, the size of the obtained RNA sequencing data is at least 100 kb. In some embodiments, the size of the obtained RNA sequencing data is at least 500 kb. In some embodiments, the size of the obtained RNA sequencing data is at least 1 megabase (Mb). In some embodiments, the size of the obtained RNA sequencing data is at least 10 Mb. In some embodiments, the size of the obtained RNA sequencing data is at least 100 Mb. In some embodiments, the size of the obtained RNA sequencing data is at least 500 Mb. In some embodiments, the size of the obtained RNA sequencing data is at least 1 gigabase (Gb). In some embodiments, the size of the obtained RNA sequencing data is at least 10 Gb. In some embodiments, the size of the obtained RNA sequencing data is at least 100 Gb. In some embodiments, the size of the obtained RNA sequencing data is at least 500 Gb.

[0130] In some embodiments, the expression data is obtained by bulk RNA sequencing. Bulk RNA sequencing may include obtaining the RNA expression level of each gene across RNA extracted from a large population of input cells (e.g., a mixture of different cell types). In some embodiments, the expression data is obtained by single-cell sequencing (e.g., scRNA-seq). Single-cell sequencing may include sequencing individual cells.

[0131] In some embodiments, the bulk sequencing data comprises at least 1 million reads, at least 5 million reads, at least 10 million reads, at least 20 million reads, at least 50 million reads, or at least 100 million reads. In some embodiments, the bulk sequencing data comprises between 1 million and 5 million reads, between 3 million and 10 million reads, between 5 million and 20 million reads, between 10 million and 50 million reads, between 30 million and 100 million reads, or between 1 million and 100 million reads (or any number of reads in between).

[0132] In some embodiments, the expression data comprises next generation sequencing (NGS) data.

[0133] RNA expression data (e.g., showing RNA expression levels) for multiple genes can be used for any of the methods or compositions described herein. The number of genes that can be examined can be up to and including all genes of the subject. In some embodiments, RNA expression levels can be determined for all genes of the subject. As non-limiting examples, 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, 16 or more, 17 or more, 18 or more, 19 or more, 20 or more, 21 or more, 22 or more, 23 or more, 24 or more, 25 or more, 26 or more, 27 or more, 28 or more, 29 or more, 30 or more, 35 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, 225 or more, 250 or more, 275 or more, or 300 or more genes can be used for any evaluation described herein. As another non-limiting example of a set, the RNA expression data may include RNA expression data for at least 5, at least 10, at least 15, at least 20, at least 25, at least 35, at least 50, at least 75, at least 100, at least 500, at least 1000, or at least 1500 genes selected from Table 2 or Table 3.

[0134] In some embodiments, the RNA expression data is obtained by accessing the RNA expression data from at least one computer storage medium on which the RNA expression data is stored. Additionally or alternatively, in some embodiments, the RNA expression data may be received from one or more sources via any suitable type of communication network. For example, in some embodiments, the RNA expression data may be received from a server (e.g., an SFTP server, or Illumina BaseSpace).

[0135] The RNA expression data obtained may be in any suitable format, as aspects of the technology described herein are not limited in this respect. For example, in some embodiments, the RNA expression data may be obtained in a text-based file (e.g., FASTQ, FASTA, BAM, or SAM format). In some embodiments, the file in which the sequencing data is stored may contain a quality score of the sequencing data. In some embodiments, the file in which the sequencing data is stored may contain sequence identifier information.

[0136] The RNA expression data, in some embodiments, includes RNA expression levels. The RNA expression levels can be detected by detecting the products of gene expression, such as mRNA and / or protein. In some embodiments, the RNA expression levels are determined by detecting the levels of mRNA in the sample. As used herein, the term "determining" or "detecting" may include assessing the presence, absence, quantity and / or amount (which may be an effective amount) of a substance, including inducing a qualitative or quantitative concentration level of such a substance in a sample, or otherwise assessing the value and / or classification of such a substance in a sample derived from a subject.

[0137] 32 illustrates an exemplary process 3200 for processing sequencing data to obtain RNA expression data from the sequencing data. Process 3200 can be performed by any suitable computing device or devices, as aspects of the technology described herein are not limited in this respect. For example, process 3200 can be performed by a computing device portion of a sequencing device. In other embodiments, process 3200 can be performed by one or more computing devices external to the sequencing device.

[0138] Process 3200 begins with activity 3201, where sequencing data is obtained from a biological sample obtained from a subject. The sequencing data is obtained by any suitable method, for example, using any of the methods described herein, including in the section entitled "Biological Samples."

[0139] In some embodiments, the sequencing data obtained at activity 3201 includes RNA-seq data. In some embodiments, the biological sample includes blood or tissue. In some embodiments, the biological sample includes one or more tumor cells.

[0140] Process 3200 then proceeds to activity 3203, where the sequencing data obtained in activity 3201 is normalized to units of transcripts per million kilobases (TPM). Normalization can be performed in any suitable manner using any suitable software. For example, in some embodiments, TPM normalization can be performed according to the techniques described in Wagner et al. (Theory Biosci. (2012) 131:281-285), which is incorporated herein by reference in its entirety. In some embodiments, TPM normalization can be performed using a software package, such as, for example, the gcrma package. Aspects of the gcrma package are described in Wu J, Gentry RlwcfJMJ (2021), "gcrma: Background Adjustment Using Sequence Information. R package version 2.66.0.", which is incorporated herein by reference in its entirety. In some embodiments, the RNA expression level in TPM for a particular gene is calculated using the following formula:

[0141]

number

[0142] Process 3200 then proceeds to activity 3205, where the RNA expression levels in TPM (determined in activity 3203) may be log transformed. Process 3200 is exemplary and variations exist. For example, in some embodiments, one or both of activities 3203 and 3205 may be omitted. Thus, in some embodiments, the RNA expression levels may not be normalized to units of transcripts per million, but may instead be converted to another type of unit (e.g., reads per million kilobases (RPKM) or fragments per million kilobases (FPKM) or any other suitable unit). Additionally or alternatively, in some embodiments, the log transformation may be omitted. Alternatively, in some embodiments, no transformation may be applied, or one or more other transformations may be applied in place of the log transformation.

[0143] The RNA expression data obtained by process 3200 may include sequence data generated by a sequencing protocol (e.g., a series of nucleotides in a nucleic acid molecule identified by next generation sequencing, Sanger sequencing, etc.) as well as information contained therein (e.g., information indicative of source, tissue type, etc.), which may also take into account information that can be inferred or determined from the sequence data. In some embodiments, the expression data obtained by process 3200 may include information contained in a FASTA file, descriptions and / or quality scores contained in a FASTQ file, aligned positions contained in a BAM file, and / or any other suitable information obtained from any suitable file.

[0144] Post-Mapping Processing The second expression level of the gene of biological sample can be used as the input for any suitable downstream technique of processing expression data.Examples of downstream processing techniques include, but are not limited to, applying quality control techniques to the second expression level, using the second expression level to associate biological sample with cohorts, using the second expression level to determine the tumor microenvironment of subject, using the expression level to perform cell analysis, and using the expression level to select therapeutic agents for subject.

[0145] In some embodiments, the second expression level of the gene of the biological sample is used as an input for applying one or more quality control techniques to the expression level. Methods for applying quality control techniques to the expression level are known, for example, as described in International Patent Application No. PCT / IB2020 / 000928, filed July 3, 2020, and published February 18, 2021 as International Publication WO2021 / 028726, the entire contents of which are incorporated herein by reference.

[0146] In some embodiments, the second expression level of the gene of biological sample is used as the input for associating biological sample with cohort.The method of associating biological sample with cohort is known, for example, as described in International Patent Application No. PCT / US2018 / 037008, filed on June 12, 2018, and published on December 20, 2018 as International Publication WO2018 / 231762, the entire contents of which are incorporated herein by reference.

[0147] In some embodiments, the second expression level of the gene of the biological sample is used as the input for determining the tumor microenvironment of the subject.The method for determining the tumor microenvironment of the subject is known, for example, as described in International Patent Application No. PCT / US2018 / 037017, filed on June 12, 2018, and published on December 20, 2018 as International Publication WO2018 / 231771, the entire contents of which are incorporated herein by reference.

[0148] In some embodiments, the second expression level of the gene of the biological sample is used as an input for performing cellular analysis. Methods for performing cellular analysis are known, for example, as described in International Patent Application No. PCT / US2021 / 022155, filed March 12, 2021, and published September 16, 2021 as International Publication WO2021 / 183917, the entire contents of which are incorporated herein by reference.

[0149] In some embodiments, the second expression level of the gene in the biological sample is used as an input for selecting a therapeutic agent for the subject.Methods for selecting a therapeutic agent for the subject are known, for example, as described in International Patent Application No. PCT / US2018 / 037008, filed on June 12, 2018, and published on December 20, 2018 as International Publication WO2018 / 231762, the entire contents of which are incorporated herein by reference.

[0150] Anti-cancer therapy Aspects of the present disclosure relate to a method of treating a subject having (or suspected of having, or at risk of having) cancer by administering to the subject a cancer therapeutic agent selected using a second expression level obtained by the method described herein. In some embodiments, the method includes administering to the subject one or more (e.g., one, two, three, four, five or more) therapeutic agents. In some embodiments, the therapeutic agent (or agent) administered to the subject is selected from small molecules, peptides, nucleic acids, radioisotopes, cells (e.g., CAR T cells, etc.), and combinations thereof. Examples of therapeutic agents include chemotherapeutic agents (e.g., cytotoxic agents, etc.), immunotherapeutic agents (e.g., immune checkpoint inhibitors such as PD-1 inhibitors, PD-L1 inhibitors, etc.), antibodies (e.g., anti-HER2 antibodies), cell therapy (e.g., CAT T cell therapy), gene silencing therapy (e.g., interfering RNA, CRISPR, etc.), antibody-drug conjugates (ADC), and combinations thereof. In some embodiments, the subject is administered an effective amount of the therapeutic agent. As used herein, "effective amount" refers to the amount of each active agent required to confer a therapeutic effect on a subject, alone or in combination with one or more other active agents. Effective amounts will vary, as recognized by those skilled in the art, depending on the particular condition being treated, the severity of the condition, individual patient parameters including age, physical condition, size, sex and weight, duration of treatment, nature of concomitant therapy (if any), specific route of administration and similar factors within the knowledge and expertise of medical personnel. These factors are well known to those skilled in the art and can be addressed using no more than routine experimentation. It is generally preferred that the maximum dose of each component or combination thereof, i.e., the highest safe dose according to sound medical judgment, is used. However, it will be understood by those skilled in the art that a patient may require a lower dose or tolerated dose for medical reasons, psychological reasons, or virtually any other reason.

[0151] Empirical considerations such as the half-life of therapeutic compounds generally contribute to the determination of dosage.For example, antibodies compatible with the human immune system, such as humanized or fully human antibodies, can be used to extend the half-life of antibodies and prevent them from being attacked by the host's immune system.The frequency of administration can be determined and adjusted over the course of therapy, and is generally (but not necessarily) based on the treatment and / or suppression and / or improvement and / or delay of cancer.Alternatively, sustained continuous release formulations of anti-cancer therapeutics may be appropriate.Various formulations and devices for achieving sustained release are known.

[0152] In some embodiments, the dosage for the anti-cancer therapeutic agent described herein can be empirically determined in an individual who has been administered one or more doses of the anti-cancer therapeutic agent. The individual can be administered an increasing dosage of the anti-cancer therapeutic agent. To evaluate the efficacy of the administered anti-cancer therapeutic agent, one or more aspects of the cancer (such as, for example, tumor microenvironment, tumor formation, tumor growth, or TME type) can be analyzed.

[0153] In general, for administration of any of the anti-cancer antibodies described herein, an initial candidate dosage may be about 2 mg / kg. For purposes of this disclosure, a typical daily dosage may range anywhere from about 0.1 μg / kg to 3 μg / kg to 30 μg / kg to 300 μg / kg to 3 mg / kg, 30 mg / kg to 100 mg / kg or more, depending on the factors mentioned above. For repeated administration over several days or more, depending on the condition, treatment is sustained until a desired suppression or amelioration of symptoms occurs, or until a therapeutic level sufficient to alleviate the cancer, or one or more symptoms thereof, is achieved. An exemplary dosing regimen includes administering an initial dose of about 2 mg / kg, followed by a weekly maintenance dose of about 1 mg / kg of the antibody, or a maintenance dose of about 1 mg / kg every two weeks. However, other dosage regimens may be useful depending on the pattern of pharmacokinetic decay the practitioner (e.g., physician) wishes to achieve. For example, dosing one to four times per week is contemplated. In some embodiments, dosages ranging from about 3 μg / mg to about 2 mg / kg (such as about 3 μg / mg, about 10 μg / mg, about 30 μg / mg, about 100 μg / mg, about 300 μg / mg, about 1 mg / kg, and about 2 mg / kg) may be used. In some embodiments, the frequency of administration is once every week, every two weeks, every four weeks, every five weeks, every six weeks, every seven weeks, every eight weeks, every nine weeks, or every ten weeks; or once every month, every two months, or every three months, or more. The progress of this therapy can be monitored by conventional techniques and assays and / or by monitoring the GC TME types described herein. The administration regimen (including the therapeutic agent used) may vary over time.

[0154] If the anti-cancer therapeutic agent is not an antibody, it can be administered at a ratio of about 0.1-300 mg per kg of patient body weight, divided into 1-3 doses, or as disclosed herein. In some embodiments, for an adult patient of normal body weight, a dose ranging from about 0.3-5.00 mg / kg can be administered. The particular dosage regimen, e.g., dose, timing, and / or repetition, will depend on the particular subject and their individual medical history, as well as the characteristics of the individual agent (such as the half-life of the agent, and other considerations as known).

[0155] For purposes of this disclosure, the appropriate dosage of an anti-cancer therapeutic will depend on the particular anti-cancer therapeutic (or composition thereof) used, the type and severity of the cancer, whether the anti-cancer therapeutic is administered for prophylactic or therapeutic purposes, previous therapy, the patient's medical history and response to the anti-cancer therapeutic, and the discretion of the attending physician. Typically, a clinician will administer an anti-cancer therapeutic, such as an antibody, until a dosage is reached that achieves the desired result.

[0156] Administration of the anti-cancer therapeutic agent may be continuous or intermittent, depending, for example, on the physiological condition of the recipient, whether the purpose of administration is therapeutic or prophylactic, and other factors known to those skilled in the art. Administration of the anti-cancer therapeutic agent (e.g., an anti-cancer antibody) may be essentially continuous over a preselected period of time, or may be in a series of spaced doses, for example, either before, during, or after the onset of cancer.

[0157] As used herein, the term "treatment" refers to the application or administration of a composition containing one or more active agents to a subject with cancer, a symptom of cancer, or a predisposition to cancer, for the purpose of curing, healing, alleviating, palliating, altering, treating, ameliorating, improving, or affecting the cancer or one or more symptoms of cancer, or a predisposition to cancer.

[0158] Alleviating cancer includes delaying the onset or progression of disease, or reducing disease severity. Alleviating disease does not necessarily require a curative outcome. As used herein, "delaying" the onset of disease (e.g., cancer) means to postpone, hinder, slow down, retard, stabilize, and / or postpone the progression of disease. This delay may be for a variable length of time, depending on the history of the disease and / or individual being treated. A method of "delaying" or alleviating the onset of disease, or delaying the onset of disease, is a method that reduces the probability of developing one or more symptoms of disease in a given time frame and / or reduces the severity of symptoms in a given time frame, compared to not using the method. Such comparisons are typically based on clinical trials, using a sufficient number of subjects to obtain statistically significant results.

[0159] "Onset" or "progression" of a disease refers to early signs of disease and / or subsequent progression of the disease. Onset of a disease can be detected and evaluated using known clinical techniques. Alternatively, or in addition to known clinical techniques, onset of a disease can be detected and evaluated based on other criteria. However, onset also refers to progression that may be undetectable. For the purposes of this disclosure, onset or progression refers to the biological course of a condition. "Onset" includes development, recurrence, and onset. As used herein, "onset" or "onset" of cancer includes initial onset and / or recurrence.

[0160] Examples of antibody anti-cancer drugs include, but are not limited to, alemtuzumab (Campath), trastuzumab (Herceptin), ibritumomab tiuxetan (Zevalin), brentuximab vedotin (Adcetris), ado-trastuzumab emtansine (Kadcyla), blinatumomab (Blincyto), bevacizumab (Avastin), cetuximab (Erbitux), ipilimumab (Yervoy), nivolumab (Opdivo), pembrolizumab (Keytruda), atezolizumab (Tecentriq), avelumab (Bavencio), durvalumab (Imfinzi), and panitumumab (Vectibix).

[0161] Examples of immunotherapies include, but are not limited to, PD-1 or PD-L1 inhibitors, CTLA-4 inhibitors, adoptive cell transfer, therapeutic cancer vaccines, oncolytic virus therapy, T-cell therapy, and immune checkpoint inhibitors.

[0162] Examples of radiation therapy include, but are not limited to, ionizing radiation, gamma radiation, neutron beam radiotherapy, electron beam radiotherapy, proton therapy, brachytherapy, systemic radioisotopes, and radiosensitizers.

[0163] Examples of surgical therapies include, but are not limited to, curative surgery (eg, tumor removal surgery), preventative surgery, laparoscopic surgery, and laser surgery.

[0164] Examples of chemotherapeutic agents include, but are not limited to, R-CHOP, carboplatin or cisplatin, docetaxel, gemcitabine, nab-paclitaxel, paclitaxel, pemetrexed, and vinorelbine. Further examples of chemotherapy agents include, but are not limited to, platinum agents such as carboplatin, oxaliplatin, cisplatin, nedaplatin, satraplatin, lobaplatin, triplatin, tetranitrate, picoplatin, prolindac, aroplatin, and other derivatives; topoisomerase I inhibitors such as camptothecin, topotecan, irinotecan / SN38, rubitecan, belotecan, and other derivatives; etoposide (VP-16), daunorubicin, doxorubicin agents (e.g., Topoisomerase II inhibitors such as doxorubicin, doxorubicin hydrochloride, doxorubicin analogs, or liposomal doxorubicin and its salts or analogs), mitoxantrone, aclarubicin, epirubicin, idarubicin, amrubicin, amsacrine, pirarubicin, valrubicin, zorubicin, teniposide and other derivatives; folic acid family (methotrexate, pemetrexed, raltitrexed, aminopterin, and its relatives or derivatives), etc. antimetabolites; purine antagonists (thioguanine, fludarabine, cladribine, 6-mercaptopurine, pentostatin, clofarabine, and their related or derivatives) and pyrimidine antagonists (cytarabine, floxuridine, azacitidine, tegafur, carmoful, capacitabine, gemcitabine, hydroxyurea, 5-fluorouracil (5FU), and their related or derivatives); nitrogen mustards (e.g., cyclophosphamide, melphalan, chloramphenicol, Alkylating agents such as bucill, mechlorethamine, ifosfamide, mechlorethamine, trofosfamide, prednimustine, bendamustine, uramustine, estramustine, and related or derivatives thereof; nitrosoureas (e.g., carmustine, lomustine, semustine, fotemustine, nimustine, ranimustine, streptozocin, and related or derivatives thereof); triazenes (e.g., dacarbazine, altretamine, temozolomide, and related or derivatives thereof);alkylsulfonates (e.g., busulfan, mannosulfan, treosulfan, and related or derivatives thereof); procarbazine; mitobronitol, and aziridines (e.g., carboquone, triaziquone, thiotepa, triethylenemalamine, and related or derivatives thereof); antibiotics such as hydroxyurea, anthracyclines (e.g., doxorubicin, daunorubicin, epirubicin, and related or derivatives thereof); anthracenediones (e.g., mitoxantrone, and related or derivatives thereof); Streptomycete antibiotics (e.g., bleomycin, mitomycin C, actinomycin, and plicamycin); and ultraviolet light.

[0165] Computer implementation An exemplary implementation of a computer system 3300 that can be used in connection with any of the embodiments of the technology described herein (e.g., such as the method of FIG. 3) is shown in FIG. 33. The computer system 3300 includes one or more articles of manufacture including one or more processors 3310 and non-transitory computer-readable storage media (e.g., memory 3320 and one or more non-volatile storage media 3330). The processor 3310 can control the writing and reading of data to and from the memory 3320 and the non-volatile storage device 3330 in any suitable manner, as aspects of the technology described herein are not limited to any particular technology for writing or reading data. To perform any of the functions described herein, the processor 3310 can execute one or more processor-executable instructions stored in one or more non-transitory computer-readable storage media (e.g., memory 3320), which can act as a non-transitory computer-readable storage medium that stores processor-executable instructions for execution by the processor 3310.

[0166] The computing device 3300 may also include a network input / output (I / O) interface 3340 through which the computing device can communicate with other computing devices (e.g., over a network), and may also include one or more user I / O interfaces 3350 through which the computing device provides output to and receives input from a user. The user I / O interfaces may include devices such as a keyboard, a mouse, a microphone, a display device (e.g., a monitor or touch screen), a speaker, a camera, and / or various types of I / O devices.

[0167] The above embodiments can be implemented in any of several ways. For example, the embodiments can be implemented using hardware, software, or a combination thereof. When implemented in software, the software code can be executed on any suitable processor (e.g., a microprocessor) or collection of processors, whether provided in a single computing device or distributed among multiple computing devices. It should be understood that any component or collection of components performing the above functions can be generally considered as one or more controllers that control the functions discussed above. The one or more controllers can be implemented in several ways, such as with dedicated hardware or with general purpose hardware (e.g., one or more processors) programmed with microcode or software to perform the functions described above.

[0168] In this regard, it should be understood that one implementation of the embodiments described herein includes at least one computer-readable storage medium (e.g., RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital versatile disk (DVD) or other optical disk storage, magnetic cassette, magnetic tape, magnetic disk storage or other magnetic storage device, or other tangible non-transitory computer-readable storage medium) encoded with a computer program (i.e., a plurality of executable instructions) that, when executed on one or more processors, performs the above-discussed functions of one or more embodiments. The computer-readable medium may be transportable such that the program stored thereon can be loaded onto any computing device to perform aspects of the technology discussed herein. Furthermore, it should be understood that references to a computer program that, when executed, performs any of the functions discussed above are not limited to application programs running on a host computer. Rather, the terms computer program and software are used herein in a generic sense to refer to any type of computer code (e.g., application software, firmware, microcode, or any other form of computer instructions) that can be used to program one or more processors to perform aspects of the technology discussed herein.

[0169] The foregoing description of the implementations provides illustration and description, but is not intended to be exhaustive or to limit the implementations to the precise form disclosed. Modifications and variations are possible in light of the above teachings or may be acquired from practice of the implementations. In other implementations, the methods depicted in these figures may include some operations, different operations, operations in a different order, and / or additional operations. Moreover, non-dependent blocks may be executed simultaneously.

[0170] It will be apparent that the exemplary aspects described above can be implemented in many different forms of software, firmware, and hardware in the implementation illustrated in the figures. Furthermore, certain parts of the implementation can be implemented as a "module" that performs one or more functions. The module may include hardware, such as a processor, an application-specific integrated circuit (ASIC), or a field-programmable gate array (FPGA), or a combination of hardware and software. EXAMPLES

[0171] Example 1: Batch Effect RNA-seq quantitatively measures gene expression across the entire genome, with higher expression values ​​corresponding to more abundant mRNAs in the sample. This linearity is a key property of RNA quantification assays and is responsible for the high (>80%) intra-sample correlation across different platforms. This inter-platform concordance of expression levels has been previously shown for qPCR / TaqMan, DNA microarrays (e.g., HuGene2) and different RNA-seq modifications (SOLID) by comparison of the log ratios between the expression levels of selected genes derived from the same sample. In the comparison, the majority of gene points (ratios) were proportional and were as follows: y = a * x (where the coefficient "a" depends on the platform pair being compared). Nearly all individual RNA expression evaluation platforms (e.g., SOLID, ribo-Zero, EC, Nugen) correlated with qPCR evaluation, thereby supporting the idea of ​​gene expression linear comparability across different methods, including polyA and EC sequencing. Notably, the linearity was more evident for a selection of protein-coding genes.

[0172] The absolute expression values ​​of genes profiled using the same protocol will vary depending on the tissue preservation method (microarray and total RNA-seq). Furthermore, if the samples are sequenced by another protocol, the absolute values ​​will change, a problem known as batch effect. Normalization, which is the adjustment of the overall characteristics of the measurements for each individual sample, does not remove the batch effect. Furthermore, the direct cause of the batch effect is technical differences; therefore, removing these technical differences does not affect biological variability. Gene expression in specimens evaluated using different GEP protocols (e.g., polyA RNA-seq, EC RNA-seq, microarray) will differ due to batch effects; however, the relative expression values ​​for all genes in comparison with each other will generally still be similar (i.e., high intra-sample multigene Pearson correlation within RNA-seq and high Spearman correlation across any platform). Although the absolute values ​​generated by different protocols may vary substantially, many genes are linearly correlated across different protocols.

[0173] Previously described batch effect correlation algorithms have been developed to neutralize batch effects between samples across large cohorts, however, these techniques are generally not suitable for batch correction of expression data obtained from individual samples.

[0174] Example 2: Single sample mapping Genetic selection This example describes the ability to apply a linear model that maps expression data from a single biological sample sequenced using a first protocol (e.g., FFPE tissue sequenced by EC RNA-seq) to reference expression data (e.g., expression data for a patient cohort) obtained from a biological sample sequenced using a protocol different from the first protocol (e.g., FF tissue sequenced by polyA RNA-seq). The performance of the algorithm described herein was improved by training it with paired samples sequenced using two different protocols, allowing data derived from the two protocols to be combined and analyzed.

[0175] Briefly, RNA transcripts per million (TPM) normalization was performed within a set of transcripts (gene isoforms) selected according to their biological type or their biological family using GENCODE v23 transcriptome annotation. For TPM normalization, all transcripts of non-coding biological types were excluded, as previously performed in The Cancer Genome Atlas (TGCA) mRNA Analysis Pipeline for FPKM. Histone-coding and mitochondrial gene transcripts were also excluded due to unequal enrichment with different RNA extraction methods, e.g., polyA vs. total RNA. The resulting set of genes, retained for TPM normalization and expression quantification, contained a set of 1,899 genes that were cancer-specific, immune-related, clinically and scientifically relevant for cancer (i.e., clinical biomarkers and genes that can be used for further processing, e.g., single sample gene set enrichment analysis (ssGSEA) and cellular analysis techniques), selected as the most relevant targets for projection from one protocol to another.

[0176] Mapping of some genes from one protocol to another may be affected by technical or biological issues. For example, some genes may not cross with the probes used for EC, and other genes may have transcripts with low annotation or reference sequence quality (e.g., low transcriptional support level, partially unknown coding sequence, etc.). There are gene families that are lost during the polyA sequencing protocol in contrast to the total RNA or EC protocol, which can be explained by specific polyadenylation (e.g., ubiquitin-specific peptidase 17-like family, Speedy / RINGO cell cycle regulator family, taste 2 receptor family, and some olfactory receptors). Also, the expression of TCR and BCR coding genes annotated in the transcriptome as corresponding to V, D or J regions cannot be properly quantified without specific alignment tools such as MiXCR. Furthermore, direct measurements of polyA length in HeLa cell lines were obtained for more than 4,000 genes. Genes with polyA length below the mean and differing by more than one standard deviation from the mean were considered to have short polyA tails. The 190 genes with the above problems were included in the target gene set of 1,899 genes (300 in total) along with 271 additional genes that often had low expression levels around the noise level measured across polyA or EC (e.g., Agilent SureSelect V7+UTR) protocols, either together or separately.

[0177] A total of four groups of genes (listed in Table 1 below) were obtained for further analysis. Tables 2 and 3 provide examples of genes in the BMG and BMGEP groups described in Table 1.

[0178] [Table 1]

[0179] Single gene mapping To explore the feasibility of creating a batch correction algorithm using paired samples sequenced with polyA or EC RNA-seq, we acquired a publicly available cohort, MET500, containing paired samples of diverse cancer types (Figure 5). In total, it contained 320 paired samples derived from the same sample that were sequenced with both polyA RNA-seq and the Agilent SureSelect V4 EC protocol.

[0180] For the MET500 cohort, PCA showed a clear separation between the expression data generated by the different protocols (Figure 6). The absolute values ​​differed for most genes; however, for many of them high Pearson correlation values ​​were observed between protocols (representative example, Figure 7).

[0181] Overall, 297 of the 320 samples passed the implemented quality control steps. 92 paired samples were selected as a holdout set to perform validation comparisons, and the remaining sample set was used to train a single-gene model (e.g., single-gene mapping, as shown in Figures 2A-B). A brief description of the single-gene model is provided below.

[0182] Given a predictor p, the linear regression model is y=W0+w1x1+...+w p x p Predict the response y by

[0183] The model fitting procedure yields a vector of coefficients w. For example, ordinary least squares (OLS) estimates are obtained by minimizing the sum of squared residuals. However, OLS often performs poorly in both prediction and interpretation. Penalization techniques are used to improve OLS. Lasso and ridge regression are penalized least squares methods that impose l1 and l2 penalties on the regression coefficients, respectively.

[0184] For expression data projection from one sequencing protocol to another, y is the projected expression and x is a vector of predictors. For the above inter-platform concordance of expression levels, if most gene points (ratios) follow a linear dependency between different platforms, a linear regression model with the formula y=w0+w1x1 (where x1 is the target gene expression in EC and y is its projection onto polyA) may be useful.

[0185] We used a machine learning tool called ElasticNet, which uses the following formula:

[0186]

number

[0187] In some embodiments, a version of ElasticNet named ElasticNetCV was used, which provides an internal cross-validation estimator that can be used to search for specialized model parameters (i.e., α and l1 ratios) with greater computational efficiency compared to standard estimators.

[0188] An ElasticNetCV regression model was used to auto-tune the parameters and the concordance correlation coefficient (CCC) was used to measure whether the algorithm accurately overcomes the batch effect between the two different techniques. A linear model (also called a "transformation") was then applied to "correct" (e.g., map) the expression values ​​in the holdout set. UMAP projections performed on the all-gene (AG) group showed that the algorithm effectively overcomes the overall batch effect while preserving unique tissue gene expression patterns (Figure 8).

[0189] Next, the correction performance of the algorithm across the biologically meaningful gene (BMG) group. The CCC values ​​for over 1518 genes were above 0.75, indicating the robust performance of the developed single-gene model (Figure 9). Thus, this type of model can be used to combine cohorts. Furthermore, individual samples can be mapped from one protocol to the expression distribution of another protocol by applying the correction.

[0190] Next, the reproducibility of gene signatures after correction was investigated. First, the values ​​of representative gene signatures (e.g., as described by US Patent Application Publication No. 2020-0273543, entitled "SYSTEMS AND METHODS FOR GENERATING, VISUALIZING AND CLASSIFYING MOLECULAR FUNCTIONAL PROFILES," the entire contents of which are incorporated herein by reference) were calculated using ssGSEA. The initial and corrected values ​​across paired polyA and EC samples were compared using CCC (polyA vs. EC-before correction and polyA vs. EC-after correction). The CCC values ​​for the majority of gene signatures before correction were greater than 90%, and increased slightly after correction (Figure 10 and Figure 11).

[0191] A comparison was then performed for Kassandra analysis (e.g., as described in U.S. Patent Application No. 17 / 200,492, filed March 12, 2021, entitled "SYSTEMS AND METHODS FOR DECONVOLUTION OF EXPRESSION DATA," which is incorporated by reference in its entirety). CCC values ​​improved significantly across all major cell types (FIG. 12), with the best results in CD4 T cells (FIG. 13).

[0192] Multigene mapping To develop a multigene model (e.g., the multigene mapping shown in Figures 2C-2D), Pearson correlation was calculated within the BMG group on TCGA expression data, including different cancer types. Figure 14 shows a representative example of highly correlated genes with Pearson correlation values ​​above 0.7 for both polyA and EC samples. Then, for each gene of interest, up to 50 most correlated genes were selected (e.g., by Pearson correlation of RNA expression levels), which were then used to build a multigene linear model. Briefly, the gene of interest and its correlated genes were used to train the multigene model. An ElasticNetCV regression model was used to auto-tune the parameters, and the concordance correlation coefficient (CCC) was used to measure whether the algorithm accurately overcomes the batch effect between the two different techniques. A transformation was then applied to "correct" (e.g., map) the expression values ​​in the holdout set. CCC values ​​were observed to be higher within the individual genes analyzed (Figure 14, second row), and improved gene-wise CCC was observed across BMGs (Figure 15) compared to single-gene mapping techniques.

[0193] Example 3: Comparisons using cohort-based adjustments To evaluate the effectiveness of the developed algorithms, we compared them against current batch correction techniques: PCA-based correction, MNN-based correction, and ComBat. The comparative analysis is given below.

[0194] PCA Principal component analysis (PCA) was performed by inverting the transformed data back to the original space after removing one or more of the most significant principal components (PCs) (Figure 16). Specifically, a matrix containing the eigenvectors can be used to obtain the PCs: PC=VX (where PC are the principal components, X are the original expression data (polyA and EC merged into one data frame) and V is the matrix containing the eigenvectors).

[0195] Thus, the inverted data after removing some PCs is given by the following formula: V T PC=X This can be achieved by solving

[0196] Results on the training data showed that the biological diversity of the expression data decreased with increasing number of PCs removed (Figures 17A-17B: column 1). Also, upon PC removal, the EC and polyA cohorts merge together and project into the same space, but are still not comparable to both the original EC and polyA expression. Thus, an attempt was made to identify a matrix by multiplication that results in the transformation of EC expression into the space of the original polyA expression (Figures 17A-17B: column 2). The results showed an improvement of the CCC performed gene-wise when removing the first PC (Figures 17A-17B: column 3).

[0197] We then performed transformation of the hold-out samples using the same PCs and the same multiplication matrix obtained from the training samples (Figures 18A-18B). The results showed a decrease in gene-wise CCC of the transformed data compared to the original expression. Thus, for the newly arrived samples, we were unable to transfer expression values ​​from EC to polyA using the training-precomputed PCs and transition matrix.

[0198] MNN-based correction Next, a method based on the detection of mutual neighbors (MNN) was compared with the single sample mapping technique. In this approach, MNN pairs represent a common population structure, which can be used to estimate the batch-corrected values. To implement this method, each sample from the holdout-EC set was taken separately (one by one) and added to the training-EC set, and the new set was then matched with the training-polyA set. This method utilizing MNNs can be described by the following steps shown in Figure 19: 1) Take one sample from the holdout-EC set; 2) Add this sample to the training-EC set to obtain a "dummy set"; 3) Match the "dummy set" with the training-polyA expression; 4) Select only the holdout samples from the transformed expression and add them to the set of "MNN transformed" samples.

[0199] The complete "MNN-transformed" set of samples was then compared to the holdout-polyA cohort. PCA projection showed that the transformed dataset did not perfectly match the polyA expression (Figure 20). In terms of CCC values, MNN-based batch correction also failed to outperform the single-sample mapping technique (Figure 21).

[0200] COMBAT Adjustment Finally, we compared the single-sample mapping technique with another well-known batch correction tool, ComBat. ComBat could not be used "out of the box" in the technique to pre-train a model and then utilize it for newly appearing single samples. Therefore, we tried the same strategy as that applied for the MNN algorithm - add the hold-out-EC samples one by one to the training-EC expression, and then merge this new data frame with the training-polyA expression (Figure 19).

[0201] The performance of both methods was evaluated by calculating the CCC for expression values ​​before and after correction. The single-sample mapping technique showed significantly higher CCC values ​​and outperformed ComBat in this test (Figure 22). PCA also showed that the transformed expression of ComBat was projected onto a different space compared to both the EC and polyA holdout data (Figure 23).

[0202] Compare Models Using various methods described in previous sections, we unified EC and polyA expression across four predefined groups of genes (e.g., Table 1) and compared their gene-wise CCC values ​​calculated on a holdout set of MET500 samples (Figure 24). Single- and multi-gene linear models showed better performance (more than 75% of genes exhibiting a CCC above 0.8) compared to the original data and other methods. Therefore, these two models were selected for further evaluation on laboratory data.

[0203] Example 4: Single sample mapping and cohort identification A model was created for the Agilent SureSelect V7+UTR protocol. In total, 88 paired samples from the same tissue section were subjected to different sample processing and sequencing procedures. FF samples were sequenced using the polyA protocol, while in-house prepared FFPE samples were sequenced using the EC protocol Agilent V7+UTR. In total, 64 paired samples were used to train the ElasticNetCV linear model (one for each gene), and the remaining 24 samples were used for the holdout data set.

[0204] When these models were applied, the batch effect dropped significantly, such that pairs of polyA and "corrected" EC samples began to group together according to the PCA projection (Figures 25-27). Also, the intra-sample correlation increased dramatically (from an average of about 85% to an average of 95%) (Figures 28 and 29). Looking at the BMG group, 1,416 of the 1,900 genes had a CCC higher than 0.75 after correction (1,292 genes were above 0.8), and 1,695 genes had a CCC higher than 0.50 (Figures 30 and 31).

[0205] Kassandra analysis was also performed to calculate CCC values ​​in the major cell types for both the predicted and validation polyA expression sets. Figure 32 shows a slight decrease in all cases except for the "fibroblast" group, where the CCC values ​​increased significantly after correction.

[0206]

Table 2-1

[0207]

Table 2-2

[0208]

Table 2-3

[0209]

Table 2-4

[0210]

Table 2-5

[0211]

Table 2-6

[0212]

Table 2-7

[0213]

Table 2-8

[0214]

Table 2-9

[0215]

Table 2-10

[0216]

Table 2-11

[0217]

Table 2-12

[0218]

Table 2-13

[0219]

Table 2-14

[0220]

Table 2-15

[0221]

Table 2-16

[0222]

Table 2-17

[0223]

Table 2-18

[0224]

Table 2-19

[0225]

Table 2-20

[0226]

Table 2-21

[0227]

Table 2-22

[0228]

Table 2-23

[0229]

Table 2-24

[0230]

Table 2-25

[0231]

Table 2-26

[0232]

Table 2-27

[0233]

Table 2-28

[0234]

Table 2-29

[0235]

Table 2-30

[0236]

Table 2-31

[0237]

Table 2-32

[0238]

Table 2-33

[0239]

Table 2-34

[0240]

Table 2-35

[0241]

Table 2-36

[0242]

Table 2-37

[0243]

Table 2-38

[0244]

Table 2-39

[0245]

Table 2-40

[0246]

Table 2-41

[0247]

Table 2-42

[0248]

Table 2-43

[0249]

Table 2-44

[0250]

Table 2-45

[0251]

Table 2-46

[0252]

Table 2-47

[0253]

Table 2-48

[0254]

Table 2-49

[0255]

Table 2-50

[0256]

Table 2-51

[0257]

Table 2-52

[0258]

Table 2-53

[0259]

Table 2-54

[0260]

Table 2-55

[0261]

Table 2-56

[0262]

Table 2-57

[0263]

Table 2-58

[0264]

Table 2-59

[0265]

Table 2-60

[0266]

Table 2-61

[0267]

Table 2-62

[0268]

Table 2-63

[0269]

Table 2-64

[0270]

Table 2-65

[0271]

Table 2-66

[0272]

Table 2-67

[0273]

Table 2-68

[0274]

Table 2-69

[0275]

Table 2-70

[0276]

Table 2-71

[0277]

Table 2-72

[0278]

Table 2-73

[0279]

Table 2-74

[0280]

Table 2-75

[0281]

Table 2-76

[0282]

Table 2-77

[0283]

Table 2-78

[0284]

Table 2-79

[0285]

Table 2-80

[0286]

Table 2-81

[0287]

Table 2-82

[0288]

Table 2-83

[0289]

Table 2-84

[0290]

Table 2-85

[0291]

Table 2-86

[0292]

Table 2-87

[0293]

Table 2-88

[0294]

Table 2-89

[0295]

Table 2-90

[0296]

Table 2-91

[0297]

Table 2-92

[0298]

Table 2-93

[0299]

Table 2-94

[0300]

Table 2-95

[0301]

Table 2-96

[0302]

Table 2-97

[0303]

Table 3-1

[0304]

Table 3-2

[0305]

Table 3-3

[0306]

Table 3-4

[0307]

Table 3-5

[0308]

Table 3-6

[0309]

Table 3-7

[0310]

Table 3-8

[0311]

Table 3-9

[0312]

Table 3-10

[0313]

Table 3-11

[0314]

Table 3-12

[0315]

Table 3-13

[0316]

Table 3-14

[0317]

Table 3-15

[0318]

Table 3-16

[0319]

Table 3-17

[0320]

Table 3-18

[0321]

Table 3-19

[0322]

Table 3-20

[0323]

Table 3-21

[0324]

Table 3-22

[0325]

Table 3-23

[0326]

Table 3-24

[0327]

Table 3-25

[0328]

Table 3-26

[0329]

Table 3-27

[0330]

Table 3-28

[0331]

Table 3-29

[0332]

Table 3-30

[0333]

Table 3-31

[0334]

Table 3-32

[0335]

Table 3-33

[0336]

Table 3-34

[0337]

Table 3-35

[0338]

Table 3-36

[0339]

Table 3-37

[0340]

Table 3-38

[0341]

Table 3-39

[0342]

Table 3-40

[0343]

Table 3-41

[0344]

Table 3-42

[0345]

Table 3-43

[0346]

Table 3-44

[0347]

Table 3-45

[0348]

Table 3-46

[0349]

Table 3-47

[0350]

Table 3-48

[0351]

Table 3-49

[0352]

Table 3-50

[0353]

Table 3-51

[0354]

Table 3-52

[0355]

Table 3-53

[0356]

Table 3-54

[0357]

Table 3-55

[0358]

Table 3-56

[0359]

Table 3-57

[0360]

Table 3-58

[0361]

Table 3-59

[0362]

Table 3-60

[0363]

Table 3-61

[0364]

Table 3-62

[0365]

Table 3-63

[0366]

Table 3-64

[0367]

Table 3-65

[0368]

Table 3-66

[0369]

Table 3-67

[0370]

Table 3-68

[0371]

Table 3-69

[0372]

Table 3-70

[0373]

Table 3-71

[0374]

Table 3-72

[0375]

Table 3-73

[0376]

Table 3-74

[0377]

Table 3-75

[0378]

Table 3-76

[0379] [Table 3-77]

[0380] Equivalent It should be understood that once certain aspects and embodiments of the technology described in this disclosure have been thus described, various changes, modifications, and improvements will readily occur to those skilled in the art. Such changes, modifications, and improvements are intended to be within the spirit and scope of the technology described herein. For example, one skilled in the art will readily envision various other means and / or structures for performing the functions and / or obtaining the results and / or one or more advantages described herein, and each of such variations and / or modifications is deemed to be within the scope of the embodiments described herein. One skilled in the art will recognize or be able to ascertain, using no more than routine experimentation, many equivalents to the specific embodiments described herein. Thus, the embodiments are presented by way of example only, and it should be understood that, within the scope of the appended claims and their equivalents, the invention may be practiced otherwise than as specifically described. In addition, any combination of two or more of the features, systems, items, materials, kits, and / or methods described herein is also included within the scope of the present disclosure, to the extent that such features, systems, items, materials, kits, and / or methods are not mutually inconsistent.

[0381] The above embodiments can be implemented in any of several ways. One or more aspects and embodiments of the present disclosure, including the performance of a process or method, may utilize program instructions executable by a device (e.g., a computer, a processor, or other device) to execute or control the performance of the process or method. In this regard, various inventive concepts can be embodied as a computer-readable storage medium (or multiple computer-readable storage media) (e.g., a computer memory, one or more floppy disks, compact disks, optical disks, magnetic tapes, flash memories, circuit configurations in field programmable gate arrays or other semiconductor devices, or other tangible computer storage media) encoded with one or more programs that, when executed on one or more computers or other processors, execute a method that implements one or more of the various embodiments described above. The computer-readable medium or media may be transportable such that the program or programs stored thereon can be loaded onto one or more different computers or other processors to implement various of the above aspects. In some embodiments, the computer-readable medium may be a non-transitory medium.

[0382] The terms "program" or "software" are used herein in a generic sense to refer to any type of computer code or set of computer-executable instructions that can be used to program a computer or other processor to implement the various aspects described above. Furthermore, according to one aspect, it should be understood that one or more computer programs that, when executed, perform the methods of the present disclosure need not reside on a single computer or processor, but may be distributed in a modular manner among several different computers or processors to implement various aspects of the present disclosure.

[0383] Computer-executable instructions may be in many forms, such as program modules, executed by one or more computers or other devices. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform particular tasks or implement particular abstract data types. Typically the functionality of the program modules may be combined or distributed as desired in various embodiments.

[0384] Also, the data structures may be stored in the computer-readable medium in any suitable form. For ease of demonstration, the data structures may be shown to have fields that are related by location in the data structure. Such relationships may similarly be achieved by assigning storage areas for the fields locations in the computer-readable medium that convey the relationship between the fields. However, any suitable mechanism may be used to establish relationships between information in fields of the data structures, including the use of pointers, tags, or other mechanisms that establish relationships between data elements.

[0385] If implemented in software, the software code can be executed on any suitable processor or collection of processors, whether provided in a single computer or distributed among multiple computers.

[0386] Further, it should be understood that the computer may be embodied in any of a number of forms, such as, by way of non-limiting example, a rack-mounted computer, a desktop computer, a laptop computer, or a tablet computer. Furthermore, the computer may be embedded in devices not generally considered to be computers, but having suitable processing capabilities, including personal digital assistants (PDAs), smartphones, tablets, or any other suitable portable or fixed electronic device.

[0387] A computer may also have one or more input and output devices. These devices may be used, among other things, to provide a user interface. Examples of output devices that may be used to provide a user interface include a printer or display screen for visual presentation of output and a speaker or other sound generating device for audible presentation of output. Examples of input devices that may be used for a user interface include keyboards and pointing devices such as mice, touchpads, and digital tablets. As another example, a computer may receive input information by speech recognition or in other audible forms.

[0388] Such computers may be interconnected by one or more networks of any suitable form, including local area networks or enterprise networks, and wide area networks, such as an Intelligent Network (IN) or the Internet. Such networks may be based on any suitable technology and operate according to any suitable protocol, and may include wireless networks, wired networks, or fiber optic networks.

[0389] Also, as described, some aspects may be embodied as one or more methods. Actions performed as part of a method may be ordered in any suitable manner. Thus, embodiments may be constructed in which actions are performed in a different order than illustrated, and may include performing some actions simultaneously, even when shown as sequential actions in the exemplary embodiments.

[0390] All definitions provided and used herein should be understood to take precedence over dictionary definitions, definitions in documents incorporated by reference, and / or ordinary meanings of the defined terms.

[0391] The indefinite articles "a" and "an," as used in the specification and claims, should be understood to mean "at least one," unless clearly indicated to the contrary.

[0392] The phrase "and / or" as used herein and in the claims should be understood to mean "one or both" of the elements so conjoined, i.e., elements that are conjunctive in some cases and disjunctive in other cases. Multiple elements listed with "and / or" should be construed in the same manner, i.e., "one or more" of the elements so conjoined. There may be other elements present other than the element specifically identified by the "and / or" clause, whether related or unrelated to the element specifically identified. Thus, as a non-limiting example, a reference to "A and / or B," when used in conjunction with an open-ended phrase such as "including," may, in one embodiment, refer only to A (which may include elements other than B); in another embodiment, it may refer only to B (which may include elements other than A); in yet another embodiment, it may refer to both A and B (which may include other elements), etc.

[0393] The phrase "at least one," as used herein and in the claims, in reference to a list of one or more elements, should be understood to mean at least one element selected from any one or more of the elements in the list of elements, but not necessarily including at least one of each and every element specifically listed in the list of elements, and not excluding any combination of elements in the list of elements. This definition also allows for elements to be present other than the elements specifically identified in the list of elements to which the phrase "at least one" refers, whether related or unrelated to the specifically identified element. Thus, as a non-limiting example, "at least one of A and B" (or, equivalently, "at least one of A or B" or, equivalently, "at least one of A and / or B") can refer, in one embodiment, to at least one A (and can include elements other than B), which may include more than one A but no B; in another embodiment, to at least one B (and can include elements other than A), which may include more than one B but no A; in yet another embodiment, to at least one A, which may include at least one A and more than one B, which may refer to at least one B (and can include other elements), etc.

[0394] In the claims, as well as in the above specification, all transitional phrases such as "comprising," "including," "carrying," "having," "containing," "including," "holding," "consisting of," and the like, are to be understood to be open-ended, i.e., to mean "including, but not limited to." Only the transitional phrases "consisting of" and "consisting essentially of" shall be closed or semi-closed transitional phrases, respectively.

[0395] The terms "approximately," "substantially," and "about" can be used to mean, in some embodiments, within ±20% of a target value, in some embodiments, within ±10% of a target value, in some embodiments, within ±5% of a target value, and in some embodiments, within ±2% of a target value. The terms "approximately," "substantially," and "about" may include the target value.

Claims

1. 1. A method for mapping RNA expression levels of genes expressed in a biological sample obtained from a subject using a first protocol to RNA expression levels that would have been determined by a second protocol if the biological sample had been processed using the second protocol instead of the first protocol, the method comprising: (A) obtaining first RNA expression data for a set of genes expressed in a biological sample obtained from a subject, wherein the set of genes includes a first gene and a second gene, the first RNA expression data indicating a first RNA expression level of the genes in the set of genes, the first RNA expression data being predetermined by processing the biological sample using a first protocol; and (B) mapping the first RNA expression levels of genes in the set of genes to the second RNA expression levels of genes in the set of genes, where the second RNA expression levels indicate RNA expression levels that would have been determined by a second protocol if the biological sample was processed using the second protocol instead of the first protocol, where the second protocol differs from the first protocol. wherein the mapping comprises using at least one computer hardware processor that executes: obtaining, from among the first RNA expression levels, a first set of RNA expression levels comprising the first RNA expression level for the first gene and zero, one, or more first RNA expression levels for zero, one, or more other genes in the set of genes associated with the first gene, and a second set of RNA expression levels comprising the first RNA expression level for the second gene and zero, one, or more other genes in the set of genes associated with the second gene; obtaining a first transformation for estimating an RNA expression level for a first gene that would have been determined according to a second protocol from the RNA expression levels of the one or more genes determined by the first protocol, and a second transformation, different from the first transformation, for estimating an RNA expression level for a second gene that would have been determined according to the second protocol from the RNA expression levels of the one or more genes determined by the first protocol, the first and second transformations being transformations determined from training data using linear or nonlinear regression techniques; the training data includes a plurality of paired values ​​of RNA expression levels for each of at least a portion of a set of genes that includes a first and a second gene, each pair of values ​​in the plurality of paired values ​​including an RNA expression level determined by applying a first protocol to a particular biological sample and another RNA expression level determined by applying a second protocol to the particular biological sample; and determining a second RNA expression level for a first gene by applying a first transformation to the first set of RNA expression levels, and a second RNA expression level for a second gene by applying a second transformation to the second set of RNA expression levels, for inclusion in the second RNA expression level. A method comprising:

2. the set of genes includes one or more additional genes and a further set of genes related to the one or more additional genes; The mapping is obtaining, from among the first RNA expression levels, a set of RNA expression levels comprising an RNA expression level for each of at least a portion of the one or more additional genes and an RNA expression level for at least a portion of genes of a further set of genes related to the one or more additional genes; obtaining a respective transformation for estimating an RNA expression level for each of one or more additional genes that may have been determined according to a second protocol, each transformation is a transformation estimated from training data using linear or nonlinear regression techniques; the training data includes a plurality of paired values ​​of RNA expression levels for each of at least a portion of a set of genes that includes one or more additional genes, each pair of values ​​in the plurality of paired values ​​including an RNA expression level determined by applying a first protocol to a particular biological sample and another RNA expression level determined by applying a second protocol to the particular biological sample; and determining a second RNA expression level for each of at least a portion of the additional genes by applying the respective transformation to the first set of RNA expression levels for inclusion in the second RNA expression level. The method of claim 1 , comprising:

3. Before mapping, determining, for each gene of at least a subset of the set of genes, a respective transformation for extrapolating from the RNA expression level of one or more genes of the subset determined by the first protocol to an RNA expression level for each gene of the subset that would have been determined according to the second protocol. The method of claim 1 or 2, comprising:

4. A method as described in claim 1 or 2, wherein the first transformation is a linear transformation and determining the first transformation is performed using a regularized linear regression technique using training data.

5. Obtaining a first set of RNA expression levels (a) obtaining a first RNA expression level for a first gene, a second gene, and zero other RNA expression levels; or 3. The method of claim 1 or 2, comprising: (b) identifying one or more other genes associated with the first gene, optionally wherein the identification is performed using Pearson correlation.

6. The method described in claim 1 or 2, wherein the gene associated with the target gene is a gene having an RNA expression level correlated with the expression level of the target gene.

7. 3. The method of claim 1 or 2, wherein the plurality of other genes in the set of genes comprises between 2 and 100 genes that are related to the first gene.

8. The method of claim 1 or 2, wherein: (a) the biological sample comprises a blood sample or a tissue sample, optionally wherein the tissue sample comprises tumor tissue; and / or (b) the subject is a mammal, optionally wherein the subject is a human.

9. 3. The method of claim 1 or 2, wherein the first expression data and the second expression data each comprise a normalized RNA expression level, and optionally, the normalized RNA expression level is normalized to transcripts per million (TPM) units.

10. 3. The method of claim 1 or 2, wherein the first protocol comprises preserving the biological sample by formalin fixation and paraffin embedding, FFPE, technique, and / or the second protocol comprises preserving the biological sample by fresh freezing technique.

11. The first protocol includes preserving the biological sample by a formalin-fixed and paraffin-embedded (FFPE) technique, and performing exome capture RNA sequencing on the FFPE-preserved biological sample; and / or 11. The method of claim 10, wherein the second protocol comprises preserving the biological sample by fresh freezing technique and performing polyA RNA sequencing on the biological sample preserved by fresh freezing.

12. 3. The method of claim 1 or 2, further comprising generating first RNA expression data by applying a first protocol to the biological sample.

13. (a) identifying a cohort; and / or (b) selecting a cancer therapeutic for the subject using the second RNA expression level. Further comprising: Cohort identification Correlating the second RNA expression level to the RNA expression level of a particular cohort of the plurality of cohorts; and identifying the subject as a member of a particular cohort to which the second RNA expression level is associated; Optionally, the choice of cancer treatment agent is: determining an RNA expression level for a plurality of gene groups using the second RNA expression level, the RNA expression levels for the plurality of gene groups comprising an RNA expression level for each gene group in the set of genes, the set of genes comprising at least one gene group associated with cancer malignancy and at least one gene group associated with the cancer microenvironment; and Using the determined RNA expression levels of the gene groups to select cancer therapeutic agents The method of claim 1 or 2, comprising:

14. At least one computer hardware processor; and At least one computer readable storage medium storing processor executable instructions which, when executed by at least one computer hardware processor, cause the at least one computer hardware processor to perform the method of claim 1 or 2. A system including:

15. At least one computer readable storage medium storing processor executable instructions which, when executed by at least one computer hardware processor, cause the at least one computer hardware processor to perform the method of claim 1 or 2.

Images