Machine Learning Guided Approach for Studying Solvation Environments

Abstract

<div><div><div><p>Toward practical modeling of local solvation effects of any solute in any solvent, we report a static and all-quantum mechanics based cluster-continuum approach for calculating single ion solvation free energies. This approach uses a global optimization procedure to identify low energy molecular clusters with different numbers of explicit solvent molecules and then employs the Smooth Overlap for Atomic Positions (SOAP) kernel to quantify the similarity between different low energy solute environments. From these data, we use sketch-map, a non-linear dimensionality reduction algorithm, to obtain a two-dimensional visual representation of the similarity between solute environments in differently sized microsolvated clusters. Without needing either dynamics simulations or an a priori knowledge of local solvation structure of the ions, this approach can be used to calculate solvation free energies with errors within five percent of experimental measurements for most cases.</p></div></div></div>

Gd‐Ni‐Sb‐SnO2 electrocatalysts for active and selective ozone production

“…A mixed implicit/explicit solvation model is used to model the solvation energies for aqueously solvated molecules. The automated computational procedure was reported in detail in previous work 47 . CHARMM forcefield 48 were first generated and then 100 microsolvated solutes per intermediate were formed using the rigid molecular optimizer module of the ABCluster code 49,50 .…”

“…The automated computational procedure was reported in detail in previous work. 47 CHARMM forcefield 48 were first generated and then 100 microsolvated solutes per intermediate were formed using the rigid molecular optimizer module of the ABCluster code. 49,50 For each intermediate, only the average of the lowest five energy structures found were used for solvation energy analysis.…”

Beyond Continuum Solvent Models in Computational Homogeneous Catalysis

“…Moreover, the conformational space strongly increases when adding several solvent molecules, making it necessary extensive conformational sampling [10]. Recently automated methods to determine how many explicit solvent molecules need to be added in hybrid cluster-continuum schemes to capture most of the interaction between the solute and the environment have been proposed [11,12].…”

Teaching a Neural Network to Attach and Detach Electrons from Molecules

“…[3][4][5][6][7] These models tend to provide highly accurate PESs for molecules and materials with a relatively low number of degrees of freedom. [8][9][10][11] Deep neural networks (DNN) 12,13 are a particular class of ML algorithms proven to be universal function approximators. 14 These DNNs are perfectly suitable to learn a representation of the PES for molecules.…”