cs.AIOct 23, 2026

What Transfers Across Protein Pockets? Experience Reuse in Molecular Optimization for Agentic Structure-Based Drug Design

Authors: Zaifei Yang, Weiyu Chen, Yaqing Wang, James Kwok

Organizations: The Hong Kong University of Science and Technology · Hong Kong Institute of AI for Science, City University of Hong Kong · Beijing Institute of Mathematical Sciences and Applications

Read PDF
Ask AI
Cite

Copy citation

DOINot available for this paper

Abstract

LLM agents optimize a ligand in a protein pocket by editing it step by step and docking each edit. Recent agents first spend extra dockings probing the complex with controlled edits; guidance induced from these probes is reported to help more than the chemical knowledge the model already carries. The probes are then discarded with the instance that produced them, so the obvious next step is to keep them in an experience base, letting a new pocket inherit what earlier ones already paid for. Such a base is read through a key: the record properties under which past edits are pooled and a new one is looked up. Using the full log of one such run over the CrossDocked2020 test-set pockets, we ask whether its keys recur on a pocket the base has never seen and still separate improving from worsening edits. \crchange{Using} the contacts at the site being edited, the action, and the round as \crchange{a} key, the base separates them sharply. Within our analysis, this separation is largely explained by how much the edit changed the molecule's size, the action, and the round, none of which describes the target. That is not because the target is irrelevant: naming it outright separates them several times more sharply. For a pocket absent from the base, transfer must rely on a description rather than its identity. Among the exact-match descriptions we evaluate, the more specific ones rarely recur, whereas the coarser ones provide little separation; the signals that do transfer primarily describe the edit and the molecule it produced rather than the target. Under the representations and lookup scheme studied here, an experience base may be more useful for prioritizing an agent's own proposals for docking than for directly prescribing what to build.

Figures & tables

#Held equalΔ^\widehat{\Delta} (pp)95% CI
 all four actions
1nothing+53.3+53.3[+47.6,+57.9][+47.6,+57.9]
2ΔN\Delta N+2.7+2.7[−3.0,+8.3][-3.0,+8.3]
 the two Replace actions only
3action+36.1+36.1[+28.9,+42.6][+28.9,+42.6]
4action, ΔN\Delta N+10.8+10.8[+4.3,+17.5][+4.3,+17.5]
Table 1: Gap in improvement rate between records labelled improve and worsen , averaged over groups that agree on the listed properties (Eq. ( 1 )); ΔN\Delta N is held at its exact value. Intervals are 95%95\% percentile intervals over 2,0002{,}000 bootstrap resamples of the pockets. Rows are comparable within a panel, not across.
#KeyRecursω2\omega^{2}Beats ΔN\Delta N95% CILift (pp)
 context
1input ligand0%27.6%———
2pocket, action0%20.6%———
3pocket0%12.0%———
 pocket description
4pocket descriptors, all six17%11.8%35%[13,58][13,58]−5.5-5.5
Table 2: Sixteen keys, grouped by information source. Recurs is the fraction of held-out records with an exact match in the base. ω2\omega^{2} measures in-sample separation of residual outcomes (Eq. ( 2 )), corrected for group count. Predictions use the majority residual sign among matching records. Beats ΔN\Delta N is the fraction of evaluated pockets with higher accuracy than a ΔN\Delta N lookup, with a 95%95\% binomial interval; Lift is the mean accuracy difference. Identity keys (rows 1–3) cannot recur. κ0\kappa_{0} is bold; HBD/HBA denote hydrogen-bond donor/acceptor counts. Appendix C gives pocket-descriptor definitions, group counts, and evaluated pockets.
Appendix figures & tables1 asset

Supplementary material from the paper’s appendix.

Appendix

Keyω2\omega^{2}CellsBeats ΔN\Delta N95% CIPockets
the edit
action †0.9%449%[39,60][39,60]93
contact features †0.2%1452%[41,62][41,62]93
contact features, action †1.6%4657%[47,67][47,67]93
the pocket alone
residue count0.1%340%[30,50][30,50]93
Appendix
Table 3: Keys built from the pocket, alone and joined to the edit, with the edit keys and pocket identity for reference. One row is one key, measured as in Table 2 . Cells is the number of groups the key forms, and Pockets the number of held-out pockets on which Beats ΔN\Delta N can be estimated at all; it falls below 9393 only for the two keys built from all six descriptors, where few held-out records find a group the base already holds. Rows marked †\dagger also appear in Table 2 .
More in this topic

Structure-Based Drug Design

Structure-based drug design uses protein structures and binding-site geometry to find, rank, and create molecules that bind therapeutic targets.

See all 45 papers

Explore similar work

CardsList
May 30, 2026

Probe Before You Edit: Probing-Guided Molecular Optimization for LLM Agents in Structure-Based Drug Design

Structure-based drug design increasingly employs LLM agents to iteratively refine ligands against a target pocket, yet a viable ligand must satisfy two often-conflicting objectives---binding affinity and druggability---which single optimization steps rarely improve together. To quantify this difficulty, we introduce two diagnostic metrics: the first measures how often a single edit improves both objectives, and the second measures how often a gain on one objective comes with a loss on the other. Applying these diagnostics to current LLM-agent pipelines exposes a consistent failure mode: the agent performs molecular editing without knowing how the pocket-ligand complex responds to local modifications, thus rarely achieving joint improvement. Inspired by medicinal chemists, who probe the pocket-ligand complex with controlled analog edits before choosing an optimization direction, we propose PROBE, an optimization framework built around edit--response probing. PROBE first decomposes the ligand into editable sites and builds a pocket-specific site map that flags where joint gains are plausible, where the two objectives are likely in tension, and where liability substructures should be changed; it then performs controlled probe edits whose responses are distilled into an EditManual. Guided by the site map and EditManual, PROBE runs an iterative multi-agent optimization loop. On the CrossDocked2020 benchmark, PROBE achieves state-of-the-art performance and substantially mitigates the failure modes exposed by our diagnostics metrics. Codes are in https://github.com/yzf-code/PROBE/tree/main

Zaifei Yang, Weiyu Chen, Yaqing Wang +1

The Hong Kong University of Science and Technology · City University of Hong Kong · Beijing Institute of Mathematical Sciences and Applications

Molecular OptimizationMulti-Agent LLM SystemsLarge Language Model-Guided OptimizationStructure-Based Drug Design
AskPDF
Mar 16, 2026

Protein Design with Agent Rosetta: A Case Study for Specialized Scientific Agents

Large language models (LLMs) are capable of emulating reasoning and using tools, creating opportunities for autonomous agents that execute complex scientific tasks. Protein design provides a natural testbed: although machine learning (ML) methods achieve strong results, these are largely restricted to canonical amino acids and narrow objectives, leaving unfilled need for a generalist tool for broad design pipelines. We introduce Agent Rosetta, an LLM agent paired with a structured environment for operating Rosetta, the leading physics-based heteropolymer design software, capable of modeling non-canonical building blocks and geometries. Agent Rosetta iteratively refines designs to achieve user-defined objectives, combining LLM reasoning with Rosetta's generality. We evaluate Agent Rosetta on design with canonical amino acids, matching specialized models and expert baselines, and with non-canonical residues -- where ML approaches fail -- achieving comparable performance. Critically, prompt engineering alone often fails to generate Rosetta actions, demonstrating that environment design is essential for integrating LLM agents with specialized software. Our results show that properly designed environments enable LLM agents to make scientific software accessible while matching specialized tools and human experts.

Jacopo Teneggi, S. M. Bargeen A. Turzo, Tanya Marwah +4

Polymathic AI · Johns Hopkins University · Center for Computational Biology, Flatiron Institute +3

Tool-Augmented Language Model AgentsAI Agents for Scientific DiscoveryProtein DesignLLM Agent Harnesses
AskPDF
May 12, 2026

ToolMol: Evolutionary Agentic Framework for Multi-objective Drug Discovery

Advances in large language models (LLMs) have recently opened new and promising avenues for small-molecule drug discovery. Yet existing LLM-based approaches for molecular generation often suffer from high rates of invalid and low-quality ligand candidates, a result of the syntactic limitations of current models with regard to molecular strings. In this paper, we introduce ToolMol\texttt{ToolMol}, an evolutionary agentic framework for de novo drug design. ToolMol\texttt{ToolMol} combines a multi-objective genetic algorithm with an agentic LLM operator that iteratively updates the ligand population. We build a comprehensive toolbox of RDKit-backed functions that allows our agentic operator to consisently make precise ligand modifications. ToolMol\texttt{ToolMol} achieves state-of-the-art performance on multi-objective property optimization tasks, discovering drug-like and synthesizable ligands that have >10%>10\% stronger predicted binding affinity compared to existing methods, evaluated on three protein targets. ToolMol\texttt{ToolMol} ligands additionally achieve state-of-the-art results in gold-standard Absolute Binding Free Energy scores, gaining over existing methods by over 35%35\%. By studying chain-of-thought reasoning traces, we observe that tool-calling enables the model to more faithfully execute its planned modifications, efficiently exploiting the strong chemical prior knowledge in LLMs.

Andrew Y. Zhou, Sharvaree Vadgama, Sumanth Varambally +3

Department of Computer Science and Engineering, UC San Diego, La Jolla, California, United States · Department of Computer Science, Stanford University, Stanford, California · Skaggs School of Pharmacy and Pharmaceutical Sciences, UC San Diego, La Jolla, California

Drug DiscoveryAI Agents for Scientific DiscoveryMulti-Objective Evolutionary OptimizationAI-Assisted Scientific ResearchLLM Tool UseMolecular Generation+2
AskPDF