MolCrystalFlow: Molecular Crystal Structure Prediction via Flow Matching

TL;DR

MolCrystalFlow is a novel flow-based generative model that predicts molecular crystal structures by disentangling molecular and packing complexities, outperforming existing models and integrating with machine learning potentials for accelerated discovery.

Contribution

We introduce MolCrystalFlow, a flow-based model that effectively predicts molecular crystal structures by modeling geometric symmetries and integrating with ML potentials, advancing structure prediction methods.

Findings

MolCrystalFlow outperforms MOFFlow on benchmark datasets.

MolCrystalFlow achieves competitive results against Genarris.

Integration with ML potentials accelerates crystal structure prediction.

Abstract

Molecular crystal structure prediction represents a grand challenge in computational chemistry due to large sizes of constituent molecules and complex intra- and intermolecular interactions. While generative modeling has revolutionized structure discovery for molecules, inorganic solids, and metal-organic frameworks, extending such approaches to fully periodic molecular crystals is still elusive. Here, we present MolCrystalFlow, a flow-based generative model for molecular crystal structure prediction. The framework disentangles intramolecular complexity from intermolecular packing by embedding molecules as rigid bodies and jointly learning the lattice matrix, molecular orientations, and centroid positions. Centroids and orientations are represented on their native Riemannian manifolds, allowing geodesic flow construction and graph neural network operations that respects geometric symmetries. We benchmark our model against a state-of-the-art generative model (MOFFlow) for large-size periodic crystals and a rule-based structure generation method (Genarris) on two open-source molecular crystal datasets. MolCrystalFlow outperforms MOFFlow while achieving competitive performance against Genarris. We also demonstrate an integration of MolCrystalFlow model with universal machine learning potential to accelerate molecular crystal structure prediction, paving the way for data-driven generative discovery of molecular crystals.