MolSculpt: Sculpting 3D Molecular Geometries from Chemical Syntax

TL;DR

MolSculpt introduces a novel framework that integrates 1D chemical syntax with 3D molecular geometry generation using diffusion models, achieving state-of-the-art results in de novo and conditional molecule generation.

Contribution

It presents a new method that deeply incorporates chemical knowledge from 1D models into 3D geometry synthesis via learnable queries and a trainable projector.

Findings

Achieves state-of-the-art performance on GEOM-DRUGS and QM9 datasets.

Demonstrates superior 3D fidelity and stability in generated molecules.

Effectively integrates chemical syntax with 3D geometry generation.

Abstract

Generating precise 3D molecular geometries is crucial for drug discovery and material science. While prior efforts leverage 1D representations like SELFIES to ensure molecular validity, they fail to fully exploit the rich chemical knowledge entangled within 1D models, leading to a disconnect between 1D syntactic generation and 3D geometric realization. To bridge this gap, we propose MolSculpt, a novel framework that "sculpts" 3D molecular geometries from chemical syntax. MolSculpt is built upon a frozen 1D molecular foundation model and a 3D molecular diffusion model. We introduce a set of learnable queries to extract inherent chemical knowledge from the foundation model, and a trainable projector then injects this cross-modal information into the conditioning space of the diffusion model to guide the 3D geometry generation. In this way, our model deeply integrates 1D latent chemical knowledge into the 3D generation process through end-to-end optimization. Experiments demonstrate that MolSculpt achieves state-of-the-art (SOTA) performance in \textit{de novo} 3D molecule generation and conditional 3D molecule generation, showing superior 3D fidelity and stability on both the GEOM-DRUGS and QM9 datasets. Code is available at https://github.com/SakuraTroyChen/MolSculpt.