A deep learning approach to automate refinement of somatic variant calling from cancer sequencing data. (December 2018)

• Record Type: Journal Article
• Authors: Ainscough, Benjamin; Barnell, Erica; Ronning, Peter; Campbell, Katie; Wagner, Alex; Fehniger, Todd; Dunn, Gavin; Uppaluri, Ravindra; Govindan, Ramaswamy; Rohan, Thomas; Griffith, Malachi; Mardis, Elaine; Swamidass, S.; Griffith, Obi
• Is Part Of: Nature genetics. Volume 50:Number 12(2018)
• Page Start: 1735
• Page End: 1743
• Publication Date: 2018-12
• Subjects: Human genetics -- Periodicals; 576.505
• DOI: 10.1038/s41588-018-0257-y
• Languages: English
• ISSNs: 1061-4036

Abstract

Abstract Cancer genomic analysis requires accurate identification of somatic variants in sequencing data. Manual review to refine somatic variant calls is required as a final step after automated processing. However, manual variant refinement is time-consuming, costly, poorly standardized, and non-reproducible. Here, we systematized and standardized somatic variant refinement using a machine learning approach. The final model incorporates 41, 000 variants from 440 sequencing cases. This model accurately recapitulated manual refinement labels for three independent testing sets (13, 579 variants) and accurately predicted somatic variants confirmed by orthogonal validation sequencing data (212, 158 variants). The model improves on manual somatic refinement by reducing bias on calls otherwise subject to high inter-reviewer variability. A machine learning approach for refinement of somatic variant calls automates this process and reduces bias stemming from inter-reviewer variability.