Exploring Discrete Flow Matching for 3D De Novo Molecule Generation

TL;DR

This paper benchmarks discrete flow matching methods for 3D de novo molecule generation, introduces FlowMol-CTMC with state-of-the-art performance, and proposes new metrics for assessing molecule quality beyond basic chemical constraints.

Contribution

It evaluates existing discrete flow matching methods, introduces a new model FlowMol-CTMC, and develops metrics for higher-order structural assessment of generated molecules.

Findings

FlowMol-CTMC achieves state-of-the-art performance with fewer parameters.

Basic chemical constraints are satisfied, but models often produce unusual functional groups.

New metrics reveal limitations in structural diversity and quality of generated molecules.

Abstract

Deep generative models that produce novel molecular structures have the potential to facilitate chemical discovery. Flow matching is a recently proposed generative modeling framework that has achieved impressive performance on a variety of tasks including those on biomolecular structures. The seminal flow matching framework was developed only for continuous data. However, de novo molecular design tasks require generating discrete data such as atomic elements or sequences of amino acid residues. Several discrete flow matching methods have been proposed recently to address this gap. In this work we benchmark the performance of existing discrete flow matching methods for 3D de novo small molecule generation and provide explanations of their differing behavior. As a result we present FlowMol-CTMC, an open-source model that achieves state of the art performance for 3D de novo design with fewer learnable parameters than existing methods. Additionally, we propose the use of metrics that capture molecule quality beyond local chemical valency constraints and towards higher-order structural motifs. These metrics show that even though basic constraints are satisfied, the models tend to produce unusual and potentially problematic functional groups outside of the training data distribution. Code and trained models for reproducing this work are available at \url{https://github.com/dunni3/FlowMol}.