使用预训练的图形神经网络与令牌混合器作为对形态动态的几何特征
Zihan Pengmei1, Chatipat Lorpaiboon1, Spencer C Guo1
1Department of Chemistry and James Franck Institute, University of Chicago, Chicago, Illinois 60637, USA.
The Journal of chemical physics
|January 28, 2025
概括
geom2vec使用预训练的图形神经网络 (GNN) 来自动提取模拟的分子特征. 这种方法增强了分子动态的分析,减少了手工劳动和提高了强度.
科学领域:
- 计算化学是一种计算化学.
- 结构生物学中的机器学习
背景情况:
- 在分子模拟中识别关键特征是困难的,需要人工努力.
- 现有的方法往往需要特定系统的知识和广泛的调整.
研究的目的:
- 介绍geom2vec,一种使用预训练图形神经网络 (GNN) 作为通用几何特征器的新方法.
- 为了使分子动态的强大和高效的分析没有手动的特征工程.
主要方法:
- 在使用自我监督的否定目标对各种分子构造进行等价GNN的预训练.
- 使用表达式令牌混合器来解释学习的GNN表示和结构关系.
- 将GNN培训与下游任务培训脱,以提高计算效率.
主要成果:
- geom2vec生成可转移的结构表示,用于在没有微调的情况下学习分子动力学.
- 学习的表征捕捉了分子结构单元 (令牌) 之间的可解释关系.
- 该方法方便分析大型分子图形,包括全原子蛋白质表示,并降低计算成本.
结论:
- geom2vec为分子模拟中的特征提取提供了一种通用,自动化的方法.
- 这种方法显著减少了手动选择特征的需要,并提高了模拟分析的可靠性.
- 分离策略允许对复杂的分子系统进行高效的分析.
相关概念视频
02:18Fischer Projections
13.0K
Learning to draw Fischer projections of molecules and understanding their relevance plays a crucial role in the visual depiction of organic molecules. A Fischer projection is a two-dimensional projection on a planar surface to simplify the three-dimensional wedge–dash representation of molecules. This is especially helpful in the case of molecules with multiple chiral centers that can be difficult to draw. Here, all the bonds of interest are represented as horizontal or vertical lines.
13.0K
02:06Newman Projections
16.3K
Different notations are used to represent the three-dimensional structure of molecules on two-dimensional surfaces. One of the most commonly used representations is the dash-wedge formula. The dashed wedges, solid wedges, and the plane lines indicate the groups situated behind the plane, coming out of the plane, and in the plane, respectively.
The organic molecules rotate across the single bonds leading to numerous temporary three-dimensional structures of varying energy known as...
The organic molecules rotate across the single bonds leading to numerous temporary three-dimensional structures of varying energy known as...
16.3K