Sohraby Farzin, Guo Jing-Yao, Nunes-Alves Ariane
Institute of Chemistry, Technische Universität Berlin, Straße des 17. Juni 135, Berlin 10623, Germany.
J Chem Inf Model. 2025 Jan 27;65(2):589-602. doi: 10.1021/acs.jcim.4c01656. Epub 2025 Jan 7.
Machine learning (ML) is a powerful tool for the automated data analysis of molecular dynamics (MD) simulations. Recent studies showed that ML models can be used to identify protein-ligand unbinding pathways and understand the underlying mechanism. To expedite the examination of MD simulations, we constructed PathInHydro, a set of supervised ML models capable of automatically assigning unbinding pathways for the dissociation of gas molecules from [NiFe] hydrogenases, using the unbinding trajectories of CO and H from [NiFe] hydrogenase as a training set. [NiFe] hydrogenases are receiving increasing attention in biotechnology due to their high efficiency in the generation of H, which is considered by many to be the fuel of the future. However, some of these enzymes are sensitive to O and CO. Many efforts have been made to rectify this problem and generate air-stable enzymes by introducing mutations that selectively regulate the access of specific gas molecules to the catalytic site. Herein, we showcase the performance of PathInHydro for the identification of unbinding paths in different test sets, including another gas molecule and a different [NiFe] hydrogenase, which demonstrates its feasibility for the trajectory analysis of a diversity of gas molecules along enzymes with mutations and sequence differences. PathInHydro allows the user to skip time-consuming manual analysis and visual inspection, facilitating data analysis for MD simulations of ligand unbinding from [NiFe] hydrogenases. The codes and data sets are available online: https://github.com/FarzinSohraby/PathInHydro.
机器学习（ML）是用于分子动力学（MD）模拟自动化数据分析的强大工具。最近的研究表明，ML模型可用于识别蛋白质 - 配体解离途径并理解其潜在机制。为了加快MD模拟的研究，我们构建了PathInHydro，这是一组有监督的ML模型，能够利用一氧化碳（CO）和氢气（H）从[NiFe]氢化酶上的解离轨迹作为训练集，自动为气体分子从[NiFe]氢化酶上的解离分配解离途径。由于[NiFe]氢化酶在氢气生成方面具有高效率，而氢气被许多人视为未来的燃料，因此它在生物技术领域正受到越来越多的关注。然而，其中一些酶对氧气（O）和一氧化碳敏感。人们已经做出了许多努力来解决这个问题，并通过引入选择性调节特定气体分子进入催化位点的突变来生成对空气稳定的酶。在此，我们展示了PathInHydro在不同测试集（包括另一种气体分子和不同的[NiFe]氢化酶）中识别解离路径的性能，这证明了其对具有突变和序列差异的酶上多种气体分子进行轨迹分析的可行性。PathInHydro允许用户跳过耗时的手动分析和目视检查，便于对[NiFe]氢化酶配体解离的MD模拟进行数据分析。代码和数据集可在网上获取：https://github.com/FarzinSohraby/PathInHydro。
