STAMP as a topology-aware molecular language for fragment-level learning and editing
Abstract
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.
Related
- Language-Model-Based Detection of Genetic Editing in Bacteria
- Xenograft supervision and joint methylation-fragment-end encoding define a reusable molecular scoring layer for cell-free tumor DNA
- Engineering Wheat for Compound Climate Extremes: Molecular Trade-Offs, Stress Memory, and Precision Genome Editing for Yield-Stable Resilience
- Improving real-time search through deep learning-enhanced fragment ion indexing
- CrossScale-GLIO: Topology-Preserving Vision-Language Alignment of MRI and Whole-Slide Histopathology for Diffuse Glioma
BibTeX
@misc{ren2026stamp,
title = {{STAMP as a topology-aware molecular language for fragment-level learning and editing}},
author = {Yuxuan Ren and Jiatong Li and Ziqiao Meng and Yatao Bian},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.18438},
eprint = {2610.18438},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.18438},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - STAMP as a topology-aware molecular language for fragment-level learning and editing
AU - Yuxuan Ren
AU - Jiatong Li
AU - Ziqiao Meng
AU - Yatao Bian
PY - 2026
DA - 2026/10/13
PB - arXiv
DO - 10.48550/arxiv.2610.18438
UR - https://arxiv.org/abs/2610.18438
AB - 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.
ER -