Selected Publication:

SHR Neuro Cancer Cardio Lipid Metab Microb

Platero-Rochart, D; Krivobokova, T; Gastegger, M; Reibnegger, G; Sanchez-Murcia, PA.
Prediction of Enzyme Catalysis by Computing Reaction Energy Barriers via Steered QM/MM Molecular Dynamics Simulations and Machine Learning
J CHEM INF MODEL. 2023; Doi: 10.1021/acs.jcim.3c00772 [OPEN ACCESS]
Web of Science PubMed FullText FullText_MUG

 

Leading authors Med Uni Graz

Platero Rochart Daniel de Jesus

Sánchez Murcia Pedro Alejandro

Co-authors Med Uni Graz

Reibnegger Gilbert

Altmetrics:

Dimensions Citations:

Plum Analytics:

Scite (citation analytics):

Abstract:

The prediction of enzyme activity is one of the mainchallengesin catalysis. With computer-aided methods, it is possible to simulatethe reaction mechanism at the atomic level. However, these methodsare usually expensive if they are to be used on a large scale, asthey are needed for protein engineering campaigns. To alleviate thissituation, machine learning methods can help in the generation ofpredictive-decision models. Herein, we test different regression algorithmsfor the prediction of the reaction energy barrier of the rate-limitingstep of the hydrolysis of mono-(2-hydroxyethyl)terephthalic acid bythe MHETase ofIdeonella sakaiensis.As a training data set, we use steered quantum mechanics/molecularmechanics (QM/MM) molecular dynamics (MD) simulation snapshots andtheir corresponding pulling work values. We have explored three algorithmstogether with three chemical representations. As an outcome, our trainedmodels are able to predict pulling works along the steered QM/MM MDsimulations with a mean absolute error below 3 kcal mol(-1) and a score value above 0.90. More challenging is the predictionof the energy maximum with a single geometry. Whereas the use of theinitial snapshot of the QM/MM MD trajectory as input geometry yieldsa very poor prediction of the reaction energy barrier, the use ofan intermediate snapshot of the former trajectory brings the scorevalue above 0.40 with a low mean absolute error (ca. 3 kcal mol(-1)). Altogether, we have faced in this work some initialchallenges of the final goal of getting an efficient workflow forthe semiautomatic prediction of enzyme-catalyzed energy barriers andcatalytic efficiencies.

© Med Uni Graz • Imprint