PT  - JOURNAL ARTICLE
AU  - Xiaoxu Yang
AU  - Xin Xu
AU  - Martin W. Breuss
AU  - Danny Antaki
AU  - Laurel L. Ball
AU  - Changuk Chung
AU  - Chen Li
AU  - Renee D. George
AU  - Yifan Wang
AU  - Taejeoing Bae
AU  - Alexej Abyzov
AU  - Liping Wei
AU  - Jonathan Sebat
AU  - NIMH Brain Somatic Mosaicism Network
AU  - Joseph G. Gleeson
TI  - DeepMosaic: Control-independent mosaic single nucleotide variant detection using deep convolutional neural networks
AID  - 10.1101/2020.11.14.382473
DP  - 2021 Jan 01
TA  - bioRxiv
PG  - 2020.11.14.382473
4099  - http://biorxiv.org/content/early/2021/04/02/2020.11.14.382473.short
4100  - http://biorxiv.org/content/early/2021/04/02/2020.11.14.382473.full
AB  - Mosaic variants (MVs) reflect mutagenic processes during embryonic development1 and environmental exposure2, accumulate with aging, and underlie diseases such as cancer and autism3. The detection of MVs has been computationally challenging due to sparse representation in non-clonally expanded tissues. While heuristic filters and tools trained on clonally expanded MVs with high allelic fractions are proposed, they show relatively lower sensitivity and more false discoveries4–9. Here we present DeepMosaic, combining an image-based visualization module for single nucleotide MVs, and a convolutional neural networks-based classification module for control-independent MV detection. DeepMosaic achieved higher accuracy compared with existing methods on biological and simulated sequencing data, with a 96.34% (158/164) experimental validation rate. Of 932 mosaic variants detected by DeepMosaic in 16 whole genome sequenced samples, 21.89-58.58% (204/932-546/932) MVs were overlooked by other methods. Thus, DeepMosaic represents a highly accurate MV classifier that can be implemented as an alternative or complement to existing methods.Competing Interest StatementThe authors have declared no competing interest.