E(3)-equivariant models cannot learn chirality: Field-based molecular generation
This addresses a fundamental limitation in molecular generation for drug design, potentially improving safety and potency predictions.
The paper proves that E(3)-equivariant models inherently fail to capture molecular chirality, a critical property for drug efficacy and safety, and introduces a field-based representation that resolves this issue while maintaining competitive performance with existing methods.
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.