TY - JOUR

T1 - Cell type-specific interpretation of noncoding variants using deep learning-based methods

AU - Sindeeva, Maria

AU - Chekanov, Nikolay

AU - Avetisian, Manvel

AU - Shashkova, Tatiana I

AU - Baranov, Nikita

AU - Malkin, Elian

AU - Lapin, Alexander

AU - Kardymon, Olga

AU - Fishman, Veniamin

N1 - This study was performed using the infrastructure of AIRI, Artificial Intelligence Research Institute (Moscow, Russia). Preliminary analysis of ENCODE datasets was performed by Veniamin Fishman in the ICG (project No. 121031800061–7). © The Author(s) 2023. Published by Oxford University Press GigaScience.

PY - 2023/3/20

Y1 - 2023/3/20

N2 - Interpretation of noncoding genomic variants is one of the most important challenges in human genetics. Machine learning methods have emerged recently as a powerful tool to solve this problem. State-of-the-art approaches allow prediction of transcriptional and epigenetic effects caused by noncoding mutations. However, these approaches require specific experimental data for training and cannot generalize across cell types where required features were not experimentally measured. We show here that available epigenetic characteristics of human cell types are extremely sparse, limiting those approaches that rely on specific epigenetic input. We propose a new neural network architecture, DeepCT, which can learn complex interconnections of epigenetic features and infer unmeasured data from any available input. Furthermore, we show that DeepCT can learn cell type-specific properties, build biologically meaningful vector representations of cell types, and utilize these representations to generate cell type-specific predictions of the effects of noncoding variations in the human genome.

AB - Interpretation of noncoding genomic variants is one of the most important challenges in human genetics. Machine learning methods have emerged recently as a powerful tool to solve this problem. State-of-the-art approaches allow prediction of transcriptional and epigenetic effects caused by noncoding mutations. However, these approaches require specific experimental data for training and cannot generalize across cell types where required features were not experimentally measured. We show here that available epigenetic characteristics of human cell types are extremely sparse, limiting those approaches that rely on specific epigenetic input. We propose a new neural network architecture, DeepCT, which can learn complex interconnections of epigenetic features and infer unmeasured data from any available input. Furthermore, we show that DeepCT can learn cell type-specific properties, build biologically meaningful vector representations of cell types, and utilize these representations to generate cell type-specific predictions of the effects of noncoding variations in the human genome.

UR - https://www.scopus.com/record/display.uri?eid=2-s2.0-85151043505&origin=inward&txGid=baa6c487a9fe14cae64e89cc804e90ab

U2 - 10.1093/gigascience/giad015

DO - 10.1093/gigascience/giad015

M3 - Article

C2 - 36971292

VL - 12

SP - 1

EP - 11

JO - GigaScience

JF - GigaScience

SN - 2047-217X

M1 - giad015

ER -