OPUS-Rota4: a gradient-based protein side-chain modeling framework assisted by deep learning-based predictors

Abstract

Accurate protein side-chain modeling is crucial for protein folding and protein design. In the past decades, many successful methods have been proposed to address this issue. However, most of them depend on the discrete samples from the rotamer library, which may have limitations on their accuracies and usages. In this study, we report an open-source toolkit for protein side-chain modeling, named OPUS-Rota4. It consists of three modules: OPUS-RotaNN2, which predicts protein side-chain dihedral angles; OPUS-RotaCM, which measures the distance and orientation information between the side chain of different residue pairs and OPUS-Fold2, which applies the constraints derived from the first two modules to guide side-chain modeling. OPUS-Rota4 adopts the dihedral angles predicted by OPUS-RotaNN2 as its initial states, and uses OPUS-Fold2 to refine the side-chain conformation with the side-chain contact map constraints derived from OPUS-RotaCM. Therefore, we convert the side-chain modeling problem into a side-chain contact map prediction problem. OPUS-Fold2 is written in Python and TensorFlow2.4, which is user-friendly to include other differentiable energy terms. OPUS-Rota4 also provides a platform in which the side-chain conformation can be dynamically adjusted under the influence of other processes. We apply OPUS-Rota4 on 15 FM predictions submitted by AlphaFold2 on CASP14, the results show that the side chains modeled by OPUS-Rota4 are closer to their native counterparts than those predicted by AlphaFold2 (e.g. the residue-wise RMSD for all residues and core residues are 0.588 and 0.472 for AlphaFold2, and 0.535 and 0.407 for OPUS-Rota4).

OPUS-Mut: studying the effect of protein mutation through side-chain modeling

“…In OPUS-Mut, the dataset used for training is the same as that in OPUS-Rota4 (19), which contains 10024 proteins in the training set and 983 proteins in the validation set, culled from the PISCES server on February 2017 (56). Note that none of the original structures and their related mutants mentioned in the four case studies of this paper are present in the dataset that is used to train the OPUS-Mut.…”

“…However, this kind of methods is limited by the discrete rotamers in the rotamer library and the accuracy of the scoring function. Recently, some deep learning-based methods have been proposed (18, 19), which improve the accuracy of side-chain modeling by a large margin.…”

An end-to-end deep learning method for rotamer-free protein side-chain packing

“…Several machine learning (ML) methods have been proposed for the task of side-chain prediction [22, 21, 30, 31, 34, 20, 32]. One of the earliest of these methods, SIDEPro [22], attempts to learn an additive energy function over pairwise atomic distances for each side-chain rotamer.…”

“…The output densities are then compared to a rotamer database and the closest matching rotamer is selected. The most recent version of OPUS-Rota4 [31] uses a pipeline of multiple deep networks to predict side-chain coordinates. The method uses predicted side-chain dihedral angles to obtain an initial model, and then applies gradient descent on predicted distance constraints to obtain a final structure.…”

Studying protein-protein interaction through side-chain modeling method OPUS-Mut

“…However, the performance is limited by the discrete rotamers in the rotamer library and the accuracy of the scoring function [4]. With the help of deep learning techniques, some new methods have been developed, which successfully capture the local environment of each residue, and improve the accuracy of side-chain modeling by a large degree.…”