DeePROG: Deep Attention-based Model for Diseased Gene Prognosis by Fusing Multi-omics Data.

Abstract

An in-depth exploration of gene prognosis using different methodologies aids in understanding various biological regulations of genes in disease pathobiology and molecular functions. Interpreting gene functions at biological and molecular levels remains a daunting yet crucial task in domains such as drug design, personalized medicine, and next-generation diagnostics. Recent advancements in omics technologies have produced diverse heterogeneous genomic datasets like micro-array gene expression, miRNA expression, DNA sequence, 3D structures, which are significant resources for understanding the gene functions. In this paper, we have proposed a novel self-attention based deep multi-modal network, named DeePROG, for the prognosis of disease affected genes based on heterogeneous omics data. This research work uses three NCBI datasets spanning over three modalities, namely gene expression profile, the underlying DNA sequence, and the 3D protein structures. To extract useful features from each of these modalities, we have developed some context-specific deep learning models. Finally, three attention-based deep bi-modal architectures along with DeePROG have been developed for the prognosis of the underlying biomedical data. We have assessed our developed models performance in terms of Computational Assessment of Function Annotation (CAFA2) evaluation metrics, and DeePROG exhibits a significant improvement over baseline models.