2019/12/04 by Geoffroy Dubourg-Felonneau, Dubourg-Felonneau, Geoffroy, Omar Darwish +11 · 1 citation
Biochemistry, Genetics and Molecular Biology · Computer Science · #AI in cancer detection #Cancer Genomics and Diagnostics #Explainable Artificial Intelligence (XAI) #FOS: Biological sciences #FOS: Computer and information sciences #Genetics, Bioinformatics, and Biomedical Research #Genomics (q-bio.GN) #Machine Learning (cs.LG) #Machine Learning (stat.ML) #Topic Modeling
paper · pdf · doi:10.48550/arxiv.1912.02065
openalex publication_date 2019/12/04 · openalex created_date 2022/07/26 · openalex updated_date 2026/07/28
The genomic profile underlying an individual tumor can be highly informative\nin the creation of a personalized cancer treatment strategy for a given\npatient; a practice known as precision oncology. This involves next generation\nsequencing of a tumor sample and the subsequent identification of genomic\naberrations, such as somatic mutations, to provide potential candidates of\ntargeted therapy. The identification of these aberrations from sequencing noise\nand germline variant background poses a classic classification-style problem.\nThis has been previously broached with many different supervised machine\nlearning methods, including deep-learning neural networks. However, these\nneural networks have thus far not been tailored to give any indication of\nconfidence in the mutation call, meaning an oncologist could be targeting a\nmutation with a low probability of being true. To address this, we present here\na deep bayesian recurrent neural network for cancer variant calling, which\nshows no degradation in performance compared to standard neural networks. This\napproach enables greater flexibility through different priors to avoid\noverfitting to a single dataset. We will be incorporating this approach into\nsoftware for oncologists to obtain safe, robust, and statistically confident\nsomatic mutation calls for precision oncology treatment choices.\n