Track synthesis, materials and catalysis as it is published.
Search papers, labs, and topics across Lattice.
SALA resolves a long-standing challenge in computational materials by generating flexible molecular conformations, unit cell geometries, and crystal packings end to end while strictly preserving space-group symmetries. The architecture achieves this via state-context separation, evolving low-dimensional asymmetric-unit coordinates and lattice parameters while parameterizing updates through the full periodic intermolecular environment. Across a diverse benchmark of 799 crystals spanning all seven crystal systems, SALA elevates top-50 hit rates from 8.1% to 72.0% over unconstrained full-cell baselines and cuts mean packing RMSD from 4.48 Å to 1.70 Å.
Crystal structure generation jumps from an 8.1% to a 72.0% hit rate by decoupling asymmetric-unit dynamics from periodic context rather than attempting unconstrained full-cell diffusion.
Molecular crystal packing shapes properties from drug bioavailability to charge transport, yet generating realistic structures requires coordinating molecular flexibility, intermolecular interactions and symmetry. Here we introduce SALA, a flow-matching model built on state--context separation: it evolves only asymmetric-unit (ASU) atomic coordinates and symmetry-compatible lattice parameters while predicting their updates from the full periodic environment. Conditioned on molecular graphs and a space group, SALA generates molecular conformation, lattice geometry and packing end to end while preserving the specified symmetry. Across 799 held-out crystals spanning four chemical classes and all seven crystal systems, SALA achieves a 72.0\% hit rate from 50 candidates per target, compared with 8.1\% for an atomistic full-cell generative baseline. Mean best-candidate packing root-mean-square deviation decreases from 4.48 to 1.70~\AA, with the advantage widening as system size and symmetry multiplicity increase. SALA thus enables end-to-end generation of complex, symmetry-constrained molecular crystals, with potential applications in CSP, structure-based pharmaceutical polymorph design and organic molecular materials discovery.