Track genomics, neuroscience and molecular biology.
Search papers, labs, and topics across Lattice.
STAMP, a novel topology-aware molecular language, explicitly encodes fragment identity, attachment sites, and tree topology to enable fragment-level learning and precise structural editing. This representation captures motif similarity and recurring transformations, outperforming a matched SMILES encoder on activity-cliff prediction and achieving state-of-the-art performance on scaffold-preserving optimization benchmarks. STAMP's deterministic compiler allows frozen language models to directly operate on indexed fragments, significantly improving molecular editing and optimization tasks.
Molecular editing and optimization leap forward as a deterministic compiler translates frozen LLMs into executable chemical design tools operating on explicit fragment topology.
Molecular fragments provide chemically meaningful units, but existing molecular languages rarely make their locations and connections explicit for editing. Here we introduce STAMP (Sequential Tree Attachment-based Motif-Preserving representation), a topology-aware molecular language that connects fragment-level representation learning with precise structural editing. STAMP separately encodes motif identity, attachment sites and tree topology, making fragments reusable across contexts and subtrees directly accessible. Learned representations capture motif similarity and recurring transformations, and lower activity-cliff prediction error on 23 of 30 targets versus a matched SMILES encoder. Compatible subtree operations support scaffold-preserving optimization, ranking first among six methods on 15 of 23 benchmark objectives. A deterministic compiler enables frozen language models to act on indexed fragments, improving molecular editing, optimization, transformation transfer and change identification by 9 to 56 percentage points over plain SMILES. Together, these results establish an executable molecular language in which learned chemical units remain directly operable during design.