Abstract

Physical interactions of proteins play key functional roles in many important cellular processes. To understand molecular mechanisms of such functions, it is crucial to determine the structure of protein complexes. To complement experimental approaches, which usually take a considerable amount of time and resources, various computational methods have been developed for predicting the structures of protein complexes. In computational modeling, one of the challenges is to identify near-native structures from a large pool of generated models. Here, we developed a deep learning–based approach named Graph Neural Network–based DOcking decoy eValuation scorE (GNN-DOVE). To evaluate a protein docking model, GNN-DOVE extracts the interface area and represents it as a graph. The chemical properties of atoms and the inter-atom distances are used as features of nodes and edges in the graph, respectively. GNN-DOVE was trained, validated, and tested on docking models in the Dockground database and further tested on a combined dataset of Dockground and ZDOCK benchmark as well as a CAPRI scoring dataset. GNN-DOVE performed better than existing methods, including DOVE, which is our previous development that uses a convolutional neural network on voxelized structure models.

Highlights

  • Determined protein structures provide fundamental information about the physicochemical nature of the biological function of protein complexes

  • A 3D grid representation of an interface often includes voxels of void space where no atoms exist inside, which is not efficient in memory usage and may even be detrimental for accurate prediction. We address this limitation of DOVE by applying a graph neural network (GNN) (Scarselli et al, 2008; Wu et al, 2020), which has previously been successful in representing molecular properties (Duvenaud et al, 2015; Smith et al, 2017; Lim et al, 2019; Zubatyuk et al, 2019)

  • We evaluated the performance of GNN-DOVE on the Dockground dataset

Read more

Summary

Introduction

Determined protein structures provide fundamental information about the physicochemical nature of the biological function of protein complexes. With the recent advances in cryo-electron microscopy, the number of experimentally determined protein complex structures has been increasing rapidly. Various approaches were used for molecular structure representations (Venkatraman et al, 2009; Pierce et al, 2011). These include docking conformational searches, such as fast Fourier transform (Katchalski-Katzir et al, 1992; Padhorny et al, 2016), geometric hashing (Fischer et al, 1995; Venkatraman et al, 2009), and particle swarm

Methods
Results
Conclusion
Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call