DeepPASTA: deep neural network based polyadenylation site analysis

Abstract

Motivation Alternative polyadenylation (polyA) sites near the 3′ end of a pre-mRNA create multiple mRNA transcripts with different 3′ untranslated regions (3′ UTRs). The sequence elements of a 3′ UTR are essential for many biological activities such as mRNA stability, sub-cellular localization, protein translation, protein binding and translation efficiency. Moreover, numerous studies in the literature have reported the correlation between diseases and the shortening (or lengthening) of 3′ UTRs. As alternative polyA sites are common in mammalian genes, several machine learning tools have been published for predicting polyA sites from sequence data. These tools either consider limited sequence features or use relatively old algorithms for polyA site prediction. Moreover, none of the previous tools consider RNA secondary structures as a feature to predict polyA sites. Results In this paper, we propose a new deep learning model, called DeepPASTA, for predicting polyA sites from both sequence and RNA secondary structure data. The model is then extended to predict tissue-specific polyA sites. Moreover, the tool can predict the most dominant (i.e. frequently used) polyA site of a gene in a specific tissue and relative dominance when two polyA sites of the same gene are given. Our extensive experiments demonstrate that DeepPASTA signisficantly outperforms the existing tools for polyA site prediction and tissue-specific relative and absolute dominant polyA site prediction. Availability and implementation https://github.com/arefeen/DeepPASTA Supplementary information Supplementary data are available at Bioinformatics online.

Alternative polyadenylation: methods, mechanism, function, and role in cancer

“… Python 2019 https://github.com/johli/aparent [ 143 ] DeepPASTA A deep learning method to predict APA from DNA sequences and RNA secondary structure data. Python 2019 https://github.com/arefeen/DeepPASTA [ 144 ] scDAPA A tool to detect and visualize APA events from single-cell RNA-seq data. R 2019 https://scdapa.sourceforge.io/ [ 145 ] APAlyzer A bioinformatics package which can examine 3’UTR-APA, intronic APA, and gene expression changes using RNA-seq data.…”

“…It was the first tool to predict poly(A) sites from both sequence and RNA secondary structure data. In addition, this tool can predict the most dominant poly(A) site of a gene in a specific tissue and predict the relative abundance of two polyA sites of the same gene [ 144 ].…”

SCAPTURE: a deep learning-embedded pipeline that captures polyadenylation information from 3’ tag-based RNA-seq of single cells

“…2c). As a result, by combining a convolution neural network (CNN) and a recurrent neural network (RNN) for data training, DeepPASS achieved an AUC over 0.99 on test datasets, substantially higher than previously reported methods, such as DeepPASTA [19] and APARENT [20] (Fig. 2d; “methods” section).…”

“…DeepPASS was compared with other poly(A) site prediction models, such as DeepPASTA [19] and APARENT [20], to assess its performance in PAS evaluation. Total 50,000 sequences were randomly selected from the same positive and negative datasets (described in DeepPASS-fixed strategy) for comparison.…”

Aptardi predicts polyadenylation sites in sample-specific transcriptomes using high-throughput RNA sequencing and DNA sequence

“…A final category of algorithms has sought to capitalize on the wealth of research connecting specific strings of DNA nucleotides, or DNA sequence elements, to polyadenylation (see Tian and Gaber 50 for a detailed review). Most of these methods, e.g., DeepPASTA 51 , Omni-PolyA 52 , and Conv-Net 53 , deploy machine learning but conspicuously do not consider in vivo expression.…”