Abstract: The present disclosure involves ctDNA assays that interrogate many regions from a single sample with high precision and accuracy, while evaluating multiple forms of cancer-related genomic alterations including sequence mutations and structural alterations. The disclosure provides simplified yet robust methods that achieve high sensitivity and specificity by analyzing cancer genes using a limited pool of non-unique barcodes in combination with endogenous barcodes. Samples are captured and sequenced using high coverage next-generation sequencing to allow tumor-specific somatic mutations, amplifications, and translocations to be identified.

Abstract: The present disclosure pertains to techniques that leverage machine learning models to identify tumor-specific mutations through an integrated analysis of next generation sequencing data. In a particular aspect, a computer-implemented method is provided that includes generating sequence reads from one or more samples collected from the same patient, generating variant call files by analyzing the sequence reads corresponding respectively to the one or more samples, comparing variant call files to generate a list of candidate somatic variants, generating, by a classification machine learning model, scores for each of the candidate somatic variants in the list of candidate somatic variants, where the scores are generated based on a plurality of classifications generated by the classification machine learning model, determining, based on the scores, a ctDNA status for the patient, where the ctDNA status is either positive or negative, and generating a report that provides the ctDNA status for the patient.

Abstract: A machine learning system and method for somatic mutation discovery are provided that provides improved identification of tumor-specific mutations. The improved identification of tumor-specific mutations may affect discovery of alterations and therapeutic management of cancer patients.

Abstract: Many regions of genomic DNA are highly similar to other regions of the genome and thus are very difficult to capture without also capturing the similar, undesired regions. This leads to over-sequencing of regions for which there is no interest and lowers coverage of the desired regions. To minimize the capture of non-desired regions, blocking baits have been designed to prevent similar but non-desired fragments from being captured. This allows more directed sequencing of the regions of interest. Blocking baits differ from capture baits in that they have modestly different sequence that preferentially bind the non-desired DNA and do not contain a biotin or other modification so remain behind when the capture baits are selected.

Abstract: A machine learning system and method for somatic mutation discovery are provided that provides improved identification of tumor-specific mutations. The improved identification of tumor-specific mutations may affect discovery of alterations and therapeutic management of cancer patients.

Abstract: The invention provides methods for determining the MSI status of a patient by liquid biopsy with sample preparation using hybrid capture and non-unique barcodes. In certain aspects, the invention provides a method of detecting microsatellite instability (MSI). The method includes obtaining cell-free DNA (cfDNA) from a sample of blood or plasma from a patient and sequencing portions of the cfDNA to obtain sequences of a plurality of tracts of nucleotide repeats in the cfDNA. A report is provided describing an MSI status in the patient when a distribution of lengths of the plurality of tracts has peaks that deviate significantly from peaks in a reference distribution.

Abstract: Many regions of genomic DNA are highly similar to other regions of the genome and thus are very difficult to capture without also capturing the similar, undesired regions. This leads to over-sequencing of regions for which there is no interest and lowers coverage of the desired regions. To minimize the capture of non-desired regions, blocking baits have been designed to prevent similar but non-desired fragments from being captured. This allows more directed sequencing of the regions of interest. Blocking baits differ from capture baits in that they have modestly different sequence that preferentially bind the non-desired DNA and do not contain a biotin or other modification so remain behind when the capture baits are selected.

Abstract: The invention provides methods for determining the MSI status of a patient by liquid biopsy with sample preparation using hybrid capture and non-unique barcodes. In certain aspects, the invention provides a method of detecting microsatellite instability (MSI). The method includes obtaining cell-free DNA (cfDNA) from a sample of blood or plasma from a patient and sequencing portions of the cfDNA to obtain sequences of a plurality of tracts of nucleotide repeats in the cfDNA. A report is provided describing an MSI status in the patient when a distribution of lengths of the plurality of tracts has peaks that deviate significantly from peaks in a reference distribution.

Abstract: A method of evaluating a sequence variation in a sample is provided. In some embodiments, the method may involve: amplifying a nucleic acid product from an initial sample; fragmenting an amount of the nucleic acid product to produce fragments; attaching an adaptor to each end of the fragments to produce adaptor-tagged fragments; sampling no more than 10% of the tagged fragments and amplifying them; sequencing at least some of the copies of the fragments to produce a plurality of sequence reads; grouping sequence reads for copies of fragments that have the same fragmentation breakpoints; deriving a consensus sequence for each of the read groups; and aligning the consensus sequences with a reference sequence.

Abstract: Cancer immunology provides promising new avenues for cancer treatment but validation of potential neoantigens to target is costly and expensive. Analysis of MHC binding affinity, antigen processing, similarity to known antigens, predicted expression levels (as mRNA or proteins), self-similarity, and mutant allele frequency, provides screening method to identify and prioritize candidate neoantigens using sequencing data. Methods of the invention thereby save time and money by identifying the priority candidate neoantigens for further experimental validation.