iScore: a novel graph kernel-based function for scoring protein–protein docking models

Abstract

Motivation Protein complexes play critical roles in many aspects of biological functions. Three-dimensional (3D) structures of protein complexes are critical for gaining insights into structural bases of interactions and their roles in the biomolecular pathways that orchestrate key cellular processes. Because of the expense and effort associated with experimental determinations of 3D protein complex structures, computational docking has evolved as a valuable tool to predict 3D structures of biomolecular complexes. Despite recent progress, reliably distinguishing near-native docking conformations from a large number of candidate conformations, the so-called scoring problem, remains a major challenge. Results Here we present iScore, a novel approach to scoring docked conformations that combines HADDOCK energy terms with a score obtained using a graph representation of the protein–protein interfaces and a measure of evolutionary conservation. It achieves a scoring performance competitive with, or superior to, that of state-of-the-art scoring functions on two independent datasets: (i) Docking software-specific models and (ii) the CAPRI score set generated by a wide variety of docking approaches (i.e. docking software-non-specific). iScore ranks among the top scoring approaches on the CAPRI score set (13 targets) when compared with the 37 scoring groups in CAPRI. The results demonstrate the utility of combining evolutionary, topological and energetic information for scoring docked conformations. This work represents the first successful demonstration of graph kernels to protein interfaces for effective discrimination of near-native and non-native conformations of protein complexes. Availability and implementation The iScore code is freely available from Github: https://github.com/DeepRank/iScore (DOI: 10.5281/zenodo.2630567). And the docking models used are available from SBGrid: https://data.sbgrid.org/dataset/684). Supplementary information Supplementary data are available at Bioinformatics online.

Structural prediction of protein interactions and docking using conservation and coevolution

“…Another scoring function driven by conservation is GraphRank, integrated in iScore. 110 In GraphRank, interfaces are not represented as a set of individual contacts but as labeled graphs in which the nodes represent interface residues, each annotated with its PSSM, and edges encode residue contacts. GraphRank classifies interfaces as native or nonnative by comparing them with a reference set of positive and negative examples.…”

“…DockRank gives good results compared to scoring functions of the reference docking program ZDOCK, partly owing to the partner‐specific trait of their interface residue predictor. Another scoring function driven by conservation is GraphRank, integrated in iScore . In GraphRank, interfaces are not represented as a set of individual contacts but as labeled graphs in which the nodes represent interface residues, each annotated with its PSSM, and edges encode residue contacts.…”

Refinement of pairwise potentials via logistic regression to score protein‐protein interactions

“…This enables integrated processing of input data, as was exemplified by the PPI scoring function ProQDock and iScore. 31,32 The rich set of supervised and unsupervised algorithms available in ML were applied to rationalize feature selection recursively to distinguish native-like ensemble of binding modes from decoys. 33 Further, the deep three-dimensional convolutional neural net, DOVE, was trained on PPI decoy set to automate the feature extraction process.…”

“…Compared with classical scoring functions, ML‐based methods have the advantage that they do not require prior assumptions between the structural data and protein‐protein complex stability. This enables integrated processing of input data, as was exemplified by the PPI scoring function ProQDock and iScore 31,32 . The rich set of supervised and unsupervised algorithms available in ML were applied to rationalize feature selection recursively to distinguish native‐like ensemble of binding modes from decoys 33 .…”

Energy-based Graph Convolutional Networks for Scoring Protein Docking Models

“…Although quality estimation methods based on machine learning are emerging for single-protein structure prediction [4][5][6][7][8][9] , such methods are still rare for protein-complex structure prediction, i.e., protein docking 10 . Second, state-of-the-art scoring functions (for relative scoring) in protein docking are often based on machine learning with hand-engineered features, such as physical-energy terms 11,12 , statistical potentials 13,14 , and graph kernels 15 . These features, heavily relying on domain expertise, are often not specifically tailored or optimized for scoring purposes.…”