Abstract: Methods, systems, and apparatus, including computer programs, for region-ambiguous joint detection. The method can include actions of obtaining a plurality of haplotypes, generating a plurality of joint diplotype candidates, wherein at least two of the plurality of haplotypes are associated with a first region of a reference genome and at least two of the other haplotypes are associated with a second region of the reference genome, the plurality of joint diplotype candidates comprising multiple different copy number configurations, determining a posterior probability for each of the plurality of joint diplotype candidates, determining a preferred copy number configuration of the multiple different copy number configurations, determining a region-ambiguous quality score for the plurality of joint diplotype candidates in a given position, and determining whether a variant exists in at least one of the first or second region of the reference genome based on the determined region-ambiguous quality score.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer-storage media, for multi-region joint detection. In some implementations, a method for identifying variants using multi-region joint detection in genetic sample sequences includes generating a set of candidate diplotypes mapped to at least two different regions in a reference sequence; generating a set of joint diplotype candidates comprising two or more candidate diplotypes of the set of candidate diplotypes; querying, for diplotypes at one or more locations of the at least two different regions in the reference sequence, a population database comprising genetic sequences of previously sequenced organisms; determining one or more values representing the frequency of specific diplotypes occurring within the genetic sequences of previously sequenced organisms; and generating, using the one or more values, an indication that the variant of the first joint diplotype candidate is an actual variant.

Abstract: Provided is a computer-implemented method, including inputting to a trained machine learning classifier genomic information of a non-training subject that includes features from a tumor sample, wherein the trained machine learning classifier trained on features of tumor samples obtained from training subjects and their a responsiveness to checkpoint inhibition treatment and the machine-learning classifier is trained to predict responsiveness to the treatment, and generating a checkpoint inhibition responsiveness classification predictive of the subject's responding to the checkpoint inhibition with the trained machine-learning classifier, and reporting the checkpoint inhibition responsiveness classification using a graphical user interface. Also provided are a computer system for performing the method and a machine learning classifier trained by the method.

Abstract: Provided is a computer-implemented method, including inputting to a trained machine learning classifier genomic information of a non-training subject that includes features from a tumor sample, wherein the trained machine learning classifier trained on features of tumor samples obtained from training subjects and their a responsiveness to checkpoint inhibition treatment and the machine-learning classifier is trained to predict responsiveness to the treatment, and generating a checkpoint inhibition responsiveness classification predictive of the subject's responding to the checkpoint inhibition with the trained machine-learning classifier, and reporting the checkpoint inhibition responsiveness classification using a graphical user interface. Also provided are a computer system for performing the method and a machine learning classifier trained by the method.

Abstract: Computer implemented methods and computer systems are provided for estimating cancer cell fractions indicating proportions of cancer cells carrying one or more mutations of interest using one or more nucleic acid samples from a subject. The methods and systems provided herein implement processes that use a variational Bayesian mixture model to cluster initial cancer cell fractions and obtain the one or more final cancer cell fractions, the initial cancer cell fractions accounting for cancer purity and copy numbers. The disclosed methods and systems improve accuracy, validity, and reliability of tests for cancer clonality, and save time, materials, cost, and computer resources required for the tests, which can help design more affective cancer treatments.