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

Loading...
Thumbnail Image

Access rights

openAccess
CC BY
acceptedVersion

URL

Journal Title

Journal ISSN

Volume Title

A4 Artikkeli konferenssijulkaisussa

Date

Major/Subject

Mcode

Degree programme

Language

en

Pages

35

Series

13th International Conference on Learning Representations, ICLR 2025, pp. 59765-59799

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. Code is available at https://dumitrescu-alexandru.github.io/FMG-web/.

Description

Publisher Copyright: © 2025 13th International Conference on Learning Representations, ICLR 2025. All rights reserved.

Keywords

Other note

Citation

Dumitrescu, A, Korpela, D, Heinonen, M, Verma, Y, Iakovlev, V, Garg, V & Lähdesmäki, H 2025, E(3)-equivariant models cannot learn chirality: Field-based molecular generation. in 13th International Conference on Learning Representations, ICLR 2025. Curran Associates Inc., pp. 59765-59799, International Conference on Learning Representations, Singapore, Singapore, 24/04/2025. < https://openreview.net/forum?id=mXHTifc1Fn >