RT Journal Article
SR Electronic
T1 MVP: predicting pathogenicity of missense variants by deep learning
JF bioRxiv
FD Cold Spring Harbor Laboratory
SP 259390
DO 10.1101/259390
A1 Hongjian Qi
A1 Chen Chen
A1 Haicang Zhang
A1 John J. Long
A1 Wendy K. Chung
A1 Yongtao Guan
A1 Yufeng Shen
YR 2018
UL http://biorxiv.org/content/early/2018/04/02/259390.abstract
AB Accurate pathogenicity prediction of missense variants is critical to improve power in genetic studies and accurate interpretation in clinical genetic testing. Here we describe a new prediction method, MVP, which uses a deep learning approach to leverage large training data sets and many correlated predictors. Using cancer mutation hotspots and de novo germline mutations from developmental disorders for benchmarking, MVP achieved better performance in prioritizing pathogenic missense variants than previous methods.