DeepAPArice: a deep learning model for poly(A) site intelligent prediction in rice using convolutional neural network

Abstract

Polyadenylation [poly(A)] is an extremely important step in the post-transcriptional process of premRNA, which plays a key role in gene regulation, protein binding and translation efficiency. Accurate identification of poly(A) sites can help understand gene expression regulation and improve the accuracy of gene annotation. Although there have been many machine learning methods used for poly(A) site prediction, most of them only take into account a limited set of features. With the increase of the amount of genomic data and the development of computing technology, deep learning technology has been used for solving many machine learning tasks. Here, we propose DeepAPArice, a deep learning model that predicts poly(A) sites in Oryza Sativa based on sequence features and RNA secondary structures around poly(A) sites. We compared DeepAPArice with both traditional machine learning algorithms and advanced deep learning algorithms, and the results show that DeepAPArice is superior to them in terms of several indicators.