Discriminative histone imputation using chromatin accessibility

Abstract

Histone modifications (HMs) play a pivot role in various biological processes, including transcription, replication and DNA repair, significantly impacting chromatin structure. These modifications underpin the molecular mechanisms of cell-specific gene expression and complex diseases. However, annotating HMs across different cell types solely using experimental approaches is impractical due to cost and time constraints. Herein, we present dHICA (discriminative histone imputation using chromatin accessibility), a novel deep learning framework that integrates DNA sequences and chromatin accessibility data to predict multiple HM tracks. Employing the Transformer architecture alongside dilated convolutions, dHICA boasts an extensive receptive field and captures more cell-type-specific information. dHICA not only outperforms state-of-the-art baselines but also achieves superior performance in cell-specific loci and gene elements, aligning with biological expectations. Furthermore, dHICAs imputations hold significant potential for downstream applications, including chromatin state segmentation and elucidating the functional implications of SNPs. In conclusion, dHICA serves as an invaluable tool for advancing the understanding of chromatin dynamics, offering enhanced predictive capabilities and interpretability.