Towards a Classification Scheme for Inferring the Atomic Composition of Drug-like Molecules from their Quantum Derived Electronic Properties

In machine learning and molecular design, there exist two approaches: discriminative and generative. In the discriminative approach dubbed forward design, the goal is to map a set of features/molecules to their respective electronic properties. In the generative approach dubbed inverse design, a set of electronic properties is given and the goal is to find the features/molecules that have these properties. These tasks are very challenging because the chemical compound space is very large. In this study, we explore a new scheme for the inverse design of molecules based on a classification paradigm that takes as input the targeted electronic properties and output the atomic composition of the molecules (i.e. atomicity or atom counts of each type in a molecule). To test this new hypothesis, we analyzed the quantum mechanics QM7b dataset consisting of 7211 small organic molecules and 14 electronic properties. Results obtained using twenty three different classification approaches including a regularized Bayesian neural network show that it is possible to achieve detection/prediction accuracy > 90%. Even though this study uses the electronic properties of molecules as input, it can be extended to other drugs' properties such as: toxicity, binding affinity, solubility, permeability, metabolic stability, etc.

Automated summary

This study proposes a new scheme for inverse design of molecules based on a classification paradigm that outputs the atomic composition of molecules given the targeted electronic properties as input. Experimental results demonstrate that this approach can achieve high accuracy. It can be extended to other drug properties, not just electronic properties of molecules.