AgentPLM: Agentic Protein Language Models with Reasoning-Augmented Decoding for Protein Sequence Design

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Rahman, Sahil, Rahman, Maxx Richard
Natura: Preprint
Pubblicazione: 2026
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866917555999866880
author Rahman, Sahil
Rahman, Maxx Richard
author_facet Rahman, Sahil
Rahman, Maxx Richard
contents Protein language models (PLMs) are passive oracles: they generate sequences in a single forward pass with no mechanism to consult external biophysical feedback or redirect generation when a candidate violates thermodynamic or structural constraints. We introduce AgentPLM, which addresses this by equipping a pre-trained PLM with i) Reasoning-Augmented Decoding (RAD), which interleaves autoregressive generation with tool calls (ESMFold, FoldX, AutoDock Vina), and ii) Contrastive Agent Policy Optimisation (CAPO), a trajectory-level extension of direct preference optimisation that trains the policy end-to-end to learn when oracle feedback is informative rather than merely imitating high-fitness sequences. We evaluate AgentPLM on benchmark tasks spanning de novo enzyme design, antibody optimisation, thermostability, PPI interface design, and zero-shot fitness prediction with standardised oracle APIs and controlled sequence-identity splits. AgentPLM achieves state-of-the-art results with a gain in antibody top-10% hit rate over the strongest passive baseline, providing mechanistic evidence of online error correction without explicit backtracking.
format Preprint
id arxiv_https___arxiv_org_abs_2606_02386
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle AgentPLM: Agentic Protein Language Models with Reasoning-Augmented Decoding for Protein Sequence Design
Rahman, Sahil
Rahman, Maxx Richard
Artificial Intelligence
Quantitative Methods
Protein language models (PLMs) are passive oracles: they generate sequences in a single forward pass with no mechanism to consult external biophysical feedback or redirect generation when a candidate violates thermodynamic or structural constraints. We introduce AgentPLM, which addresses this by equipping a pre-trained PLM with i) Reasoning-Augmented Decoding (RAD), which interleaves autoregressive generation with tool calls (ESMFold, FoldX, AutoDock Vina), and ii) Contrastive Agent Policy Optimisation (CAPO), a trajectory-level extension of direct preference optimisation that trains the policy end-to-end to learn when oracle feedback is informative rather than merely imitating high-fitness sequences. We evaluate AgentPLM on benchmark tasks spanning de novo enzyme design, antibody optimisation, thermostability, PPI interface design, and zero-shot fitness prediction with standardised oracle APIs and controlled sequence-identity splits. AgentPLM achieves state-of-the-art results with a gain in antibody top-10% hit rate over the strongest passive baseline, providing mechanistic evidence of online error correction without explicit backtracking.
title AgentPLM: Agentic Protein Language Models with Reasoning-Augmented Decoding for Protein Sequence Design
topic Artificial Intelligence
Quantitative Methods
url https://arxiv.org/abs/2606.02386