TY - JOUR
T1 - Control-independent mosaic single nucleotide variant detection with DeepMosaic
AU - NIMH Brain Somatic Mosaicism Network
AU - Yang, Xiaoxu
AU - Xu, Xin
AU - Breuss, Martin W.
AU - Antaki, Danny
AU - Ball, Laurel L.
AU - Chung, Changuk
AU - Shen, Jiawei
AU - Li, Chen
AU - George, Renee D.
AU - Wang, Yifan
AU - Bae, Taejeong
AU - Cheng, Yuhe
AU - Abyzov, Alexej
AU - Wei, Liping
AU - Alexandrov, Ludmil B.
AU - Sebat, Jonathan L.
AU - Averbuj, Dan
AU - Roy, Subhojit
AU - Courchesne, Eric
AU - Huang, August Y.
AU - D’Gama, Alissa
AU - Dias, Caroline
AU - Walsh, Christopher A.
AU - Ganz, Javier
AU - Lodato, Michael
AU - Miller, Michael
AU - Li, Pengpeng
AU - Rodin, Rachel
AU - Hill, Robert
AU - Bizzotto, Sara
AU - Khoshkhoo, Sattar
AU - Zhou, Zinan
AU - Lee, Alice
AU - Barton, Alison
AU - Galor, Alon
AU - Chu, Chong
AU - Bohrson, Craig
AU - Gulhan, Doga
AU - Maury, Eduardo
AU - Lim, Elaine
AU - Lim, Euncheon
AU - Melloni, Giorgio
AU - Cortes, Isidro
AU - Lee, Jake
AU - Kwon, Minseok
AU - Langmead, Ben
AU - Jaffe, Andrew
AU - Paquola, Apua
AU - Weinberger, Daniel
AU - Shin, Jooheon
N1 - Publisher Copyright:
© 2023, The Author(s), under exclusive licence to Springer Nature America, Inc.
PY - 2023/6
Y1 - 2023/6
N2 - Mosaic variants (MVs) reflect mutagenic processes during embryonic development and environmental exposure, accumulate with aging and underlie diseases such as cancer and autism. The detection of noncancer MVs has been computationally challenging due to the sparse representation of nonclonally expanded MVs. Here we present DeepMosaic, combining an image-based visualization module for single nucleotide MVs and a convolutional neural network-based classification module for control-independent MV detection. DeepMosaic was trained on 180,000 simulated or experimentally assessed MVs, and was benchmarked on 619,740 simulated MVs and 530 independent biologically tested MVs from 16 genomes and 181 exomes. DeepMosaic achieved higher accuracy compared with existing methods on biological data, with a sensitivity of 0.78, specificity of 0.83 and positive predictive value of 0.96 on noncancer whole-genome sequencing data, as well as doubling the validation rate over previous best-practice methods on noncancer whole-exome sequencing data (0.43 versus 0.18). DeepMosaic represents an accurate MV classifier for noncancer samples that can be implemented as an alternative or complement to existing methods.
AB - Mosaic variants (MVs) reflect mutagenic processes during embryonic development and environmental exposure, accumulate with aging and underlie diseases such as cancer and autism. The detection of noncancer MVs has been computationally challenging due to the sparse representation of nonclonally expanded MVs. Here we present DeepMosaic, combining an image-based visualization module for single nucleotide MVs and a convolutional neural network-based classification module for control-independent MV detection. DeepMosaic was trained on 180,000 simulated or experimentally assessed MVs, and was benchmarked on 619,740 simulated MVs and 530 independent biologically tested MVs from 16 genomes and 181 exomes. DeepMosaic achieved higher accuracy compared with existing methods on biological data, with a sensitivity of 0.78, specificity of 0.83 and positive predictive value of 0.96 on noncancer whole-genome sequencing data, as well as doubling the validation rate over previous best-practice methods on noncancer whole-exome sequencing data (0.43 versus 0.18). DeepMosaic represents an accurate MV classifier for noncancer samples that can be implemented as an alternative or complement to existing methods.
UR - http://www.scopus.com/inward/record.url?scp=85145373232&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85145373232&partnerID=8YFLogxK
U2 - 10.1038/s41587-022-01559-w
DO - 10.1038/s41587-022-01559-w
M3 - Article
C2 - 36593400
AN - SCOPUS:85145373232
SN - 1087-0156
VL - 41
SP - 870
EP - 877
JO - Nature biotechnology
JF - Nature biotechnology
IS - 6
ER -