Fourier series of atomic radial distribution functions: A molecular fingerprint for machine learning models of quantum chemical properties

TL;DR

This paper introduces a Fourier series-based molecular fingerprint that is invariant, continuous, and differentiable, designed for machine learning models predicting quantum chemical properties across diverse molecules.

Contribution

The authors develop a novel molecular fingerprint using Fourier series of atomic radial distribution functions, which is invariant and suitable for machine learning applications.

Findings

Achieves a mean absolute error of 8.0 kcal/mol on molecular enthalpy predictions.

Performs comparably to Coulomb matrix descriptors in property prediction.

Validated on a large set of 134,000 organic molecules.

Abstract

We introduce a fingerprint representation of molecules based on a Fourier series of atomic radial distribution functions. This fingerprint is unique (except for chirality), continuous, and differentiable with respect to atomic coordinates and nuclear charges. It is invariant with respect to translation, rotation, and nuclear permutation, and requires no pre-conceived knowledge about chemical bonding, topology, or electronic orbitals. As such it meets many important criteria for a good molecular representation, suggesting its usefulness for machine learning models of molecular properties trained across chemical compound space. To assess the performance of this new descriptor we have trained machine learning models of molecular enthalpies of atomization for training sets with up to 10k organic molecules, drawn at random from a published set of 134k organic molecules. We validate the descriptor on all remaining molecules of the 134k set. For a training set of 5k molecules the fingerprint descriptor achieves a mean absolute error of 8.0 kcal/mol, respectively. This is slightly worse than the performance attained using the Coulomb matrix, another popular alternative, reaching 6.2 kcal/mol for the same training and test sets.