Prediction of Protein-Binding Sites in DNA Sequences

Abstract

Listen

Translate article

Many of the protein binding sites (BSs) in DNA are those by transcription factors (TFs). The identification of these TFBSs in DNA sequences is very important for further understanding of underlying gene regulatory networks. When a set of regulatory regions of co-regulated genes are compared, only the significant appearance of similar short segments is observed, and this motivates the development of specialized motif-finding (or motif-discovery) algorithms (instead of conventional sequence aligners). Although a number of algorithms have been developed, the inherent problem of many false positives has remained. With the progress of the ChIP-seq technology, etc., requirements for these algorithms have been modulated. More recently, many studies based on a variety of deep-learning techniques have been introduced.