Toward Distributed Robustness in Resistance-Aware Lead Optimization
Drug resistance often enters lead optimization after a compound has been chosen. We define distributed robustness as redistributing predicted binding support away from predefined mutable contacts while retaining wild-type feasibility and panel coverage. ResistAgent separates site-level and observed combination-level liabilities, diagnoses their structural effects, and compares a matched-budget robust objective with an average-affinity comparator. Across 10 EGFR pairs, the median reduction in hotspot dependence, ΔDep (naive − robust; positive values favor the robust objective), was +0.042; across 9 valid HIV-RT pairs, it was +0.026. Panel passing was preserved but not improved: both objectives passed in every valid pair. ABL1 yielded 2/4 pairs without a corresponding robustness gain. The objective changes the dependence organization and candidate selection. Provider comparisons, ranking benchmarks, and explicit-water contact descriptors are reported alongside the primary endpoint.
Authors
- Yu Fu (ORCID: https://orcid.org/0000-0002-5463-4215)
- Weiwei Han (ORCID: https://orcid.org/0000-0002-1931-9316)
- Silvio E. Perea (ORCID: https://orcid.org/0000-0002-5791-604X)
- Huizi Cui (ORCID: https://orcid.org/0009-0008-4831-6107)
- B Y Xiao
- Yuxuan Wang (ORCID: https://orcid.org/0009-0008-4281-2388)
- Wen Li
- Jingyi Kang
Institutions
- Jilin University (CN)
- Hunan University of Science and Engineering (CN)
- Centro de Ingeniería Genética y Biotecnología (CU)
Publication Details
- Journal
- Journal of Medicinal Chemistry
- Published
- 2026-11-18
- DOI
- https://doi.org/10.1021/acs.jmedchem.6c02107
- Primary Topic
- Computational Drug Discovery Methods
- Type
- article
- Field-Weighted Citation Impact
- 0.00
Funders
- National Natural Science Foundation of China