Energy-based models for atomic-resolution protein conformations
Yilun Du Massachusetts Institute of Technology Cambridge, MA、Joshua Meier Facebook AI Research New York, NY、Jerry Ma Facebook AI Research Menlo Park, CA、Rob Fergus Facebook AI Research、New York University New York, NY、Alexander Rives New York University New York, NY Work performed during an internship at Facebook
Abstract
We propose an energy-based model (EBM) of protein conformations that operates at atomic scale. The model is trained solely on crystallized protein data. By contrast, existing approaches for scoring conformations use energy functions that incorporate knowledge of physical principles and features that are the complex product of several decades of research and tuning. To evaluate the model, we benchmark on the rotamer recovery task, the problem of predicting the conformation of a side chain from its context within a protein structure, which has been used to evaluate energy functions for protein design. The model achieves performance close to that of the Rosetta energy function, a state-of-the-art method widely used in protein structure prediction and design. An investigation of the model’s outputs and hidden representations finds that it captures physicochemical properties relevant to protein energy.
中文速览
用原子精度直接从蛋白质晶体结构数据中学习一个能量函数，长期以来都依赖物理化学专家知识和数十年手工调参，这篇论文试图绕开这条路。研究者提出了一种基于 Transformer 架构的能量模型（energy-based model，EBM），输入蛋白质中一组原子的三维坐标和化学类型，直接输出该构象的能量标量，完全不引入人工设计的物理特征。为了训练和评估模型，他们采用"旋转异构体恢复"（rotamer recovery）任务——把蛋白质侧链的真实构象遮住，看模型能否从周围原子环境中挑出能量最低、最接近真实晶体结构的构象——并与蛋白质结构预测和设计领域的主流工具 Rosetta 对比。结果显示，纯数据驱动的深度学习模型在该任务上的表现与 Rosetta 能量函数相当接近，而且对模型隐层表示的分析表明它自发学到了氢键、疏水堆积等物理化学规律。这一结果说明，无需大量专家先验知识，神经网络也能从结构数据中涌现出有意义的蛋白质能量表示，为未来将学习型能量项与物理力场结合、推动蛋白质设计自动化打开了一扇门。
原文 arXiv:2004.13167;中英对照 + 大白话阅读 https://aha.fim.ai/paper/2004.13167v1