arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2110.02423cs.LGq-bio.BMq-bio.QM
收藏

几何Transformers在蛋白质界面接触预测中的应用

Geometric Transformers for Protein Interface Contact Prediction

University of Missouri(密苏里大学)

Alex Morehead, Chen Chen, Jianlin Cheng

摘要:

用于预测蛋白质之间的界面接触的计算方法在药物发现中受到高度追捧,因为它们可以显着提高替代方法的准确性,例如蛋白质-蛋白质对接,蛋白质功能分析工具和蛋白质生物信息学的其他计算方法。在这项工作中,我们提出了几何Transformer,一种用于旋转和平移不变蛋白质界面接触预测的新型几何演化图Transformer,封装在DeepInteract中,DeepInteract是一种端到端的预测管道。DeepInteract预测伴侣特异性蛋白质界面接触(即,蛋白质间残基-残基接触),给出两种蛋白质的3D三级结构作为输入。在严格的基准测试中,DeepInteract在第13和第14次CASP-CAPRI实验以及对接基准5的挑战性蛋白质复合物目标上分别达到了14%和1.1%的最高L/5精度(L:复合物中蛋白质单元的长度)。在这样做的过程中,DeepInteract以Geometric Transformer作为其基于图形的主干,除了与DeepInteract兼容的其他基于图形的神经网络主干之外,还优于现有的界面接触预测方法,从而验证了Geometric Transformer在学习3D蛋白质结构下游任务的丰富关系几何特征方面的有效性。

英文摘要:

Computational methods for predicting the interface contacts between proteins come highly sought after for drug discovery as they can significantly advance the accuracy of alternative approaches, such as protein-protein docking, protein function analysis tools, and other computational methods for protein bioinformatics. In this work, we present the Geometric Transformer, a novel geometry-evolving graph transformer for rotation and translation-invariant protein interface contact prediction, packaged within DeepInteract, an end-to-end prediction pipeline. DeepInteract predicts partner-specific protein interface contacts (i.e., inter-protein residue-residue contacts) given the 3D tertiary structures of two proteins as input. In rigorous benchmarks, DeepInteract, on challenging protein complex targets from the 13th and 14th CASP-CAPRI experiments as well as Docking Benchmark 5, achieves 14% and 1.1% top L/5 precision (L: length of a protein unit in a complex), respectively. In doing so, DeepInteract, with the Geometric Transformer as its graph-based backbone, outperforms existing methods for interface contact prediction in addition to other graph-based neural network backbones compatible with DeepInteract, thereby validating the effectiveness of the Geometric Transformer for learning rich relational-geometric features for downstream tasks on 3D protein structures.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑