Towards accurate machine learning predictions of properties of the P–O bond cleaving in ATP upon enzymatic hydrolysis

• Published in: Mendeleev communications Vol. 34; no. 6; pp. 776 - 779
• Main Authors: Polyakov, Igor V., Miroshnichenko, Kirill D., Mulashkina, Tatiana I., Moskovsky, Alexander A., Marchenko, Ekaterina I., Khrenova, Maria G.
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.11.2024
• Subjects: ATP hydrolysis; Laplacian of electron density; machine learning; myosin; QM/MM molecular dynamics; myosin; Laplacian of electron density; QM/MM molecular dynamics; ATP hydrolysis; machine learning
• ISSN: 0959-9436
• DOI: 10.1016/j.mencom.2024.10.003

[Display omitted] Molecular dynamic simulations using QM/MM potentials are performed for the enzyme–substrate complex of adenosine triphosphate (ATP) with the motor protein myosin. Machine learning methods are applied to a dataset consisting of the geometry parameters of the active site in the enzyme–substrate complex to predict the Laplacian of electron density at the bond critical point of the PG–O3B bond being broken in ATP. Using a gradient boosting machine learning model, a mean absolute error of 0.01 a.u. and an R2 score of 0.99 are achieved, and it is found that the PG–O3B bond length is the most important feature, contributing 2/3, while other geometry features contribute 1/3.