Transcriptional Regulation Interactions

Transcriptional Regulation Interactions tool TRI_tool is an open-accessed web service for finding transcriptional regulation interactors.

Introduction

A large part of post-genomic research is focused on the analysis of protein-protein interactions (PPIs) being central to all biological processes. Inferring PPIs involved in human transcriptional regulation (TR) is of particular interest as they are often deregulated in complex diseases and may represent valuable pharmaceutical targets.

We devised an on-line tool to analyze and predict these interactions based on sequence information only aiming to evade limitations imposed by dispersed auxiliary information, such as structural or expression data. This predictor integrates information on the pseudo-amino acid composition (PseAAC) of features that dominate PPIs. Beside to the hydrophobicity, hydrophilicity and side-chain mass, it incorporates the electron-ion interaction potential (EIIP) a descriptor of long-range interaction properties that contribute considerably to protein binding specificity.

Based on a dataset that was compiled from HIPPIE (Human Integrated Protein-Protein Interaction rEference), where sequences with more than 40% of homology similarity are eliminated, a random forest model was constructed with an average value of accuracy of 80% and AUC=0.878 for independent test sets. Compared to previous studies, our approach outperformed other models in predictive performance and algorithmic efficiency and will, therefore, facilitate the understanding of the complex cellular behaviors and organizing of large-scale data into models of cellular signaling and regulatory machinery.

Please cite
Perovic V, Sumonja N, Gemovic B, Toska E, Roberts SG, Veljkovic N. TRI_tool: a web-tool for prediction of protein-protein interactions in human transcriptional regulation. Bioinformatics. 2016. pii: btw590.