MemDis: Predicting Disordered Regions in Transmembrane Proteins.

Laszlo Dobson,Gábor E Tusnády

doi:10.3390/ijms222212270

Abstract

Transmembrane proteins (TMPs) play important roles in cells, ranging from transport processes and cell adhesion to communication. Many of these functions are mediated by intrinsically disordered regions (IDRs), flexible protein segments without a well-defined structure. Although a variety of prediction methods are available for predicting IDRs, their accuracy is very limited on TMPs due to their special physico-chemical properties. We prepared a dataset containing membrane proteins exclusively, using X-ray crystallography data. MemDis is a novel prediction method, utilizing convolutional neural network and long short-term memory networks for predicting disordered regions in TMPs. In addition to attributes commonly used in IDR predictors, we defined several TMP specific features to enhance the accuracy of our method further. MemDis achieved the highest prediction accuracy on TMP-specific dataset among other popular IDR prediction methods.

Highlights

Regions in Transmembrane Proteins.Transmembrane proteins (TMPs) are located in different membranes and they provide gates between the inner and outer side of cells or organelles
We utilized Convolutional Neural Networks (CNNs) to capture local features of the sequence represented by Position-Specific Scoring Matrix and Long Short-Term Memory (LSTM)
To realistically capture the different flavors of disorder in membrane proteins, four different models were created according to different topological regions

Summary

Introduction

Regions in Transmembrane Proteins.Transmembrane proteins (TMPs) are located in different membranes and they provide gates between the inner and outer side of cells or organelles. Around 25% of the coded proteins in the human proteome contain one or more membrane regions [1]. These segments embedded in the lipid bilayer are structurally well defined; their tail and loop regions often contain unstructured segments. Such regions are aiding various functions from providing flexible linkers to binding motifs for other molecules [2]. Intrinsically disordered regions (IDRs) are well studied in general, the currently available prediction methods have limited accuracy on membrane proteins for several reasons [3]. We utilized Convolutional Neural Networks (CNNs) to capture local features of the sequence represented by Position-Specific Scoring Matrix and Long Short-Term Memory (LSTM)

Objectives

Methods

Results