E(3)-equivariant models cannot learn chirality: Field-based molecular generation

TL;DR

This paper demonstrates that E(3)-equivariant models cannot learn molecular chirality and introduces a novel field-based representation to effectively incorporate chirality in molecular generation tasks.

Contribution

The authors prove the limitations of E(3)-equivariant models in learning chirality and propose a new field-based approach that captures chirality while maintaining competitive performance.

Findings

E(3)-equivariant models cannot learn chirality.

Field-based representation captures all molecular geometries including chirality.

Proposed model achieves competitive benchmarking results.

Abstract

Obtaining the desired effect of drugs is highly dependent on their molecular geometries. Thus, the current prevailing paradigm focuses on 3D point-cloud atom representations, utilizing graph neural network (GNN) parametrizations, with rotational symmetries baked in via E(3) invariant layers. We prove that such models must necessarily disregard chirality, a geometric property of the molecules that cannot be superimposed on their mirror image by rotation and translation. Chirality plays a key role in determining drug safety and potency. To address this glaring issue, we introduce a novel field-based representation, proposing reference rotations that replace rotational symmetry constraints. The proposed model captures all molecular geometries including chirality, while still achieving highly competitive performance with E(3)-based methods across standard benchmarking metrics.