Cancer is a genomic disease where the accumulation of genetic mutations in somatic cells contributes to tumorigenesis and metastasis. Accurate characterization of somatic mutations in tumors is crucial for personalized cancer treatments in clinical care. In collaboration with Golden Helix, Sentieon Inc. provides industry-standard tools for processing next-generation sequencing data, including tools that match the somatic variant callers MuTect and MuTect2 with over 10x speedups, scalable multithreading, no downsampling and improved determinism. Sentieon has built upon these achievements with a newly released somatic variant caller, TNscope. TNscope uses an improved mathematical model to provide substantially better accuracy over MuTect2 along with additional variant annotations. We further improve TNscope’s accuracy with a random forest machine-learning model for candidate variant filtration, resulting in a highly accurate variant caller with a single ensemble quality score.
About the Presenter
Dr. Donald Freed is currently a Bioinformatics Scientist at Sentieon, Inc. He earned his Ph.D. in Neuroscience from the Johns Hopkins School of Medicine where he investigated the role of somatic mosaic mutations in autism. He is interested in deciphering genetics variants underlying human phenotypes and accurately processing large genomic datasets.