Learning relationships between over-represented motifs in a set of DNA sequences

摘要

Finding relationships between DNA sequence motifs, such as transcription factor binding sites, is an important step to understand transcription regulation in a particular context. Current computational tools are not well adapted for discovering relationships. We have developed a software system, ModuleInducer, which integrates motif finding with the analysis of possible interactions between them in the set of related DNA sequences using inductive logic programming. Our method was tested on synthetic and two kinds of real biological data. It has been shown to perform well as a cis-regulatory module finder as well as a knowledge mining tool for ChIP-Sequencing data analysis. Our method has proven to be of high suggestive value for future research by uncovering novel motif interactions in ChIP-Seq data, missed in the original study. ModuleInducer is available at: http://induce.eecs.uottawa.ca.