Implicitly perturbed Hamiltonian as a class of versatile and general-purpose molecular representations for machine learning

Amin Alibakhshi,Bernd Hartke

doi:10.1038/s41467-022-28912-6

Abstract

Unraveling challenging problems by machine learning has recently become a hot topic in many scientific disciplines. For developing rigorous machine-learning models to study problems of interest in molecular sciences, translating molecular structures to quantitative representations as suitable machine-learning inputs play a central role. Many different molecular representations and the state-of-the-art ones, although efficient in studying numerous molecular features, still are suboptimal in many challenging cases, as discussed in the context of the present research. The main aim of the present study is to introduce the Implicitly Perturbed Hamiltonian (ImPerHam) as a class of versatile representations for more efficient machine learning of challenging problems in molecular sciences. ImPerHam representations are defined as energy attributes of the molecular Hamiltonian, implicitly perturbed by a number of hypothetic or real arbitrary solvents based on continuum solvation models. We demonstrate the outstanding performance of machine-learning models based on ImPerHam representations for three diverse and challenging cases of predicting inhibition of the CYP450 enzyme, high precision, and transferrable evaluation of non-covalent interaction energy of molecular systems, and accurately reproducing solvation free energies for large benchmark sets.

Highlights

Unraveling challenging problems by machine learning has recently become a hot topic in many scientific disciplines
We demonstrate the efficiency of the machine-learning models based on ImPerHam representations in studying three challenging and extensively required problems in molecular sciences
In addition to inhibition of a cytochrome P450 (CYP450) enzyme by machinelearning, we investigate the performance of ImPerHam representations in the machine-learning evaluation of solvation free energy as well as non-covalent interaction energy, benchmarked for large datasets

Summary

Introduction

Unraveling challenging problems by machine learning has recently become a hot topic in many scientific disciplines. This method considers a linear or non-linear dependency between molecular properties and the functional groups present in molecules The success of this defined representation in predicting many properties of chemicals for several decades[12–19] has motivated its employment for approximating potential energies of molecular ensembles[20–23], as one of the most extensively studied and, at the same time, most challenging applications of machine learning in theoretical and quantum chemistry[24,25]. An excellent review of the recent progress in employing machine learning for evaluating potential energy has been reported by Manzhos and Carrington[30] These more advanced representations, despite being more efficient compared to elementary representations like functional group numbers and types, still suffer from the limited utility in studying many challenging and complicated problems of interest

Objectives

Methods

Results