A two-stream convolutional neural network for microRNA transcription start site feature integration and identification

Abstract

MicroRNAs (miRNAs) play important roles in post-transcriptional gene regulation and phenotype development. Understanding the regulation of miRNA genes is critical to understand gene regulation. One of the challenges to study miRNA gene regulation is the lack of condition-specific annotation of miRNA transcription start sites (TSSs). Unlike protein-coding genes, miRNA TSSs can be tens of thousands of nucleotides away from the precursor miRNAs and they are hard to be detected by conventional RNA-Seq experiments. A number of studies have been attempted to computationally predict miRNA TSSs. However, high-resolution condition-specific miRNA TSS prediction remains a challenging problem. Recently, deep learning models have been successfully applied to various bioinformatics problems but have not been effectively created for condition-specific miRNA TSS prediction. Here we created a two-stream deep learning model called D-miRT for computational prediction of condition-specific miRNA TSSs (http://hulab.ucf.edu/research/projects/DmiRT/). D-miRT is a natural fit for the integration of low-resolution epigenetic features (DNase-Seq and histone modification data) and high-resolution sequence features. Compared with alternative computational models on different sets of training data, D-miRT outperformed all baseline models and demonstrated high accuracy for condition-specific miRNA TSS prediction tasks. Comparing with the most recent approaches on cell-specific miRNA TSS identification using cell lines that were unseen to the model training processes, D-miRT also showed superior performance.

Recent progress in miRNA biogenesis and decay

“…More recently, the DIANA-miRGen v4 uses CAGE and ChIP-seq data to collect information on cell type-specific miRNA promoters and provide extensive prediction of dynamic miRNA promoters across tissues and cells [ 9 ], giving a good starting point for further characterization and experimental validation. Similarly, a deep learning model called D-mirT allows for computational prediction of condition-specific miRNA TSSs based on epigenetic features and sequencing data [ 10 ]. The development of these tools to predict dynamics and tissue- and condition-specific use of miRNA promoters is a very important foundation for further understanding the functional impact of differential miRNA promoter usage in biology.…”

A Deep Learning Method for MiRNA/IsomiR Target Detection

“…Deep Learning models are infamous for being a black box when it comes to understanding the underlying features. But recent studies have focused on various strategies that can reveal the features or patterns learned by different types of machine learning models [38][39][40][41][42]. Here, two of the most popular feature identification methods, convolutional kernel analysis and input perturbation, were applied to discover important features for miRNA/isomiR-mRNA interactions [26].…”

Molecular Mechanisms of Nutrient-Mediated Regulation of MicroRNAs in Pancreatic β-cells

“…MiRNA transcription start sites (TSS) are often tissue-specific and conventional techniques for miRNA TSS identification such as 5'RACE, and RNAseq has been somewhat restricted by the quick processing of pri-miRNAs into mature miRNAs and by the fact that miRNA TSS can be located thousands of nucleotides away from the mature miRNA sequence. The quick improvement in the resolution of highthroughput sequencing technologies and the computational analysis (including deep machine learning) has more recently allowed the integration of epigenetic marks (H3K4me3, H3K9/ 14Ac, PolII) and cap analysis of gene expression (CAGE-seq) to more precisely map miRNA transcription start sites in a cellspecific manner (205)(206)(207). More recently, Liu et al (206) used global nuclear run-on sequencing [GRO-seq (208)] and precision nuclear run-on sequencing (PRO-seq [209)] that provide sharp peaks around transcription start sites and continuous signal over active transcription regions to map the TSS of 480 intergenic miRNAs in 27 different human cell lines.…”