arXiv subjects

Hanqun Cao

Publications and source records attributed to Hanqun Cao.

At least 19 recordsLinked to original sources

arXiv2026

AgentFold: Closed-Loop Agentic Search for Protein Folding Model Design

Scientific LLM agents have shown promise in literature reasoning, tool use, and experiment planning, but it remains unclear whether they can autonomously improve large, tightly coupled scientific machine-learning systems through executable code changes and computationally expensive validation. We study this question in protein folding, where progress requires coordinated architectural modifications, multi-objective evaluation, and domain-aware interpretation. We present AgentFold, a multi-agent framework that formulates folding-model development as a closed-loop search over executable code variants. Starting from ESMFold, AgentFold proposes hypotheses, implements and debugs code-level modifications, evaluates model variants, analyzes experimental outcomes, and stores both successful and failed interventions in structured memory. An MCTS-style policy allocates computational resources across high-scoring search branches. On an engineering-scale protein-folding codebase comprising more than 2,000 lines of code, AgentFold explores approximately 80 model variants using approximately 5,000 GPU-hours and 170 million LLM tokens. Under a matched computational budget, AgentFold improves the best lDDT by 7.5% over independent Codex proposals and outperforms a random-search control. Beyond model improvement, the resulting intervention traces reveal recurring empirical design patterns: stable gains tend to arise from early, soft, learnable priors and gated refinement, whereas direct geometric perturbations and geometry-conditioned feedback often destabilize training. The code and experimental resources are publicly available at https://github.com/lmqfly/AgentFold.

cs.AI↗

arXiv2026

SF-Cluster: Frustration-Guided MSA Subsampling for Alternative Protein Conformation Recovery

Deep-learning structure predictors are sensitive to their multiple sequence alignment (MSA) input, making MSA subsampling a practical route to recovering alternative conformations. Existing approaches such as AF-Cluster operate in sequence space, providing limited control over which conformational basin is sampled. We introduce SF-Cluster, which subsamples MSAs using patterns of predicted local energetic frustration, a representation largely independent of sequence similarity. Across a benchmark of 48 cases spanning fold-switching, allosteric, oligomerization-coupled, and intrinsically disordered systems, and using an AF-Cluster-style dual-reference RMSD criterion, SF-Cluster improves target-state recovery of the alternative conformation over AF-Cluster across the two-state classes, with the largest improvement observed for allosteric systems (+15.5 percentage points). The selected MSAs transfer to an architecturally distinct predictor, indicating that the conformational signal resides in MSA composition. Mechanistically, matched-depth controls show that this recovery advantage is largely explained by the effective depth of the selected subsets, which frustration-pattern selection reliably reaches. At the same time, highly frustrated residues are enriched at sites supported by deep mutational scanning and NMR two-state exchange, and frustration covariation is enriched at state-switching contacts while remaining distinct from coevolutionary coupling. Together, these results identify frustration patterns as a transferable representation for conformational prediction and position MSA subsampling as a representation-guided reweighting problem.

q-bio.BM↗

arXiv2026

AlloGen: Conformation-Selective Binder Generation with Differential State Scoring

Protein binder design has largely optimized for affinity alone, leaving conformational selectivity unaddressed: for allosteric targets such as kinases, nuclear receptors, and GPCRs, a binder that engages both active and inactive states provides no functional specificity regardless of how tightly it binds. We introduce AlloGen, a modular framework that decouples backbone generation from a learned state-selectivity scorer $Q_\theta$, an SE(3)-invariant interface graph transformer trained via a two-phase curriculum that first learns interface geometry before imposing conformational discrimination. Because $Q_\theta$ is fully differentiable and generator-agnostic, it integrates with any backbone generator as a passive reranker or an active gradient-based guide without retraining. Across a diverse benchmark of proteins spanning multiple families and conformational mechanisms, AlloGen consistently identifies binders that preferentially recognize desired structural states while rejecting alternative conformations. Experimental validation on calmodulin further demonstrates that these computational selectivity signals translate to physical molecules, yielding de novo peptides that bind the desired holo conformation while exhibiting no detectable binding to the apo state. Together, these results establish conformational selectivity as a learnable property and provide a general framework for state-selective protein binder design.

q-bio.BM↗

arXiv2026

GeoCycler: Reward-Aligned 3D Diffusion for Constraint-Conditioned Cyclic Peptide Design

Cyclic peptides are attractive therapeutic modalities because their closed-ring topology can improve stability and target specificity. However, de novo cyclic peptide design remains challenging for diffusion generators, as macrocyclization requires satisfying sparse, non-smooth, and compositional geometric constraints. Existing constraint-conditioned methods largely rely on inference-time guidance, which can steer samples toward desired closures but does not directly change the learned generative distribution. We propose GeoCycler, a reward-weighted diffusion alignment framework for training conditional latent diffusion models toward macrocyclization feasibility. GeoCycler introduces a type-gated stair reward that activates distance-based shaping only when prerequisite residue or linker types are satisfied, providing dense geometric feedback while avoiding misleading signals from chemically incompatible anchors. Together with positive-only reward weighting and replay-based stabilization, GeoCycler aligns a single generator across multiple cyclization topologies. On the LNR benchmark, GeoCycler improves pass@5 closure success over strong guidance-based baselines across stapled, head-to-tail, disulfide, and bicyclic settings. In particular, it improves head-to-tail success by 20.8 percentage points over CP-Composer while maintaining comparable amino-acid and backbone-dihedral statistics. These results suggest that training-time alignment to sparse geometric constraints is a promising alternative to relying solely on post hoc sampling-time correction for cyclic peptide generation.

cs.CE↗

arXiv2026

TD3B: Transition-Directed Discrete Diffusion for Allosteric Binder Generation

Protein function is often controlled by ligands that bias the direction of state transitions, such as agonists and antagonists, rather than stabilizing a single conformation. This is especially important for clinically relevant G protein-coupled receptors (GPCRs), where therapeutic efficacy depends on functional directionality. Structure-based design methods optimize binding to static conformations and cannot represent non-reversible, directional effects or systematically distinguish agonist from antagonist behavior. To address this gap, we introduce Transition-Directed Discrete Diffusion for Allosteric Binder Design (TD3B), a sequence-based generative framework that designs binders with specified agonist or antagonist behavior via a directional transition control objective. TD3B combines a target-aware Direction Oracle, a soft binding-affinity gate, and amortized fine-tuning of a pre-trained discrete diffusion model, enabling targeted agonist and antagonist generation decoupled from binding affinity and unattainable by equilibrium-based or inference-only guidance baselines. The code and checkpoints are available at https://huggingface.co/ChatterjeeLab/TD3B.

q-bio.BM↗

arXiv2026

CA-DEL: An Open Multi-Target, Multi-Modal Benchmark for Learning from DNA-Encoded Library Screens

The success of machine learning in drug discovery hinges on learning the relationship between a chemical structure and its biological activity. While DNA-Encoded Library (DEL) technology can generate the massive datasets required for this task, its primary signal -- sequencing read counts -- is an indirect and often noisy proxy for true molecular binding affinity. To address the scarcity of public benchmarks for developing robust models that can overcome this data challenge, we introduce CA-DEL, a multi-dimensional public benchmark featuring screens against three homologous carbonic anhydrase isoforms. While recent benchmarks like KinDEL have introduced 3D poses for kinase targets, CA-DEL distinguishes itself by focusing on the selectivity challenge among homologous Carbonic Anhydrase isoforms (CAII, CAIX, CAXII). Unlike benchmarks relying solely on noisy enrichment scores, CA-DEL integrates a rigorous validation set of experimentally determined binding affinities ($K_i$) from ChEMBL, establishing a critical Sim-to-Real evaluation paradigm: training on noisy DEL screens and testing on high-fidelity biophysical data.

q-bio.BM↗

arXiv2026

NoiseRater: Meta-Learned Noise Valuation for Diffusion Model Training

Diffusion models have achieved remarkable success across a wide range of generative tasks, yet their training paradigm largely treats injected noise as uniformly informative. In this work, we challenge this assumption and introduce NoiseRater, a meta-learning framework for instance-level noise valuation in diffusion model training. We propose a parametric noise rater that assigns importance scores to individual noise realizations conditioned on data and timestep, enabling adaptive reweighting of the training objective. The rater is trained via bilevel optimization to improve downstream validation performance after inner-loop diffusion updates. To enable efficient deployment, we further design a decoupled two-stage pipeline that transitions from soft weighting during meta-training to hard noise selection during standard training. Extensive experiments on FFHQ and ImageNet demonstrate that not all noise samples contribute equally, and that prioritizing informative noise improves both training efficiency and generation quality. Our results establish noise valuation as a complementary and previously underexplored axis for improving diffusion model training. Our code is available at: https://anonymous.4open.science/r/NoiseRater-DEB116.

cs.LG↗

arXiv2026

Proteo-R1: Reasoning Foundation Models for De Novo Protein Design

Deep learning in de novo protein design has achieved atomic-level fidelity. However, existing models remain largely non-deliberative: they directly synthesize molecular geometries without explicitly reasoning about which residues or interactions are functionally essential. As a result, design decisions are entangled with continuous sampling dynamics, limiting interpretability, controllability, and systematic reuse of biochemical knowledge. We introduce Proteo-R1, a reasoning-guided protein design framework that explicitly decouples molecular understanding from geometric generation. Proteo-R1 adopts a dual-expert architecture in which a multimodal large language model (MLLM) serves as an understanding expert, analyzing protein sequences, structures, and textual context to identify key functional residues that govern binding and specificity. These residue-level decisions are then passed as hard constraints to a separate diffusion-based generation expert, which performs conditional co-design while respecting the fixed interaction anchors. This factorization mirrors how human experts approach molecular engineering: first, reasoning about critical interactions, then optimizing geometry subject to those constraints. By operationalizing reasoning as explicit residue-level commitments rather than latent textual guidance, Proteo-R1 achieves stable, interpretable, and modular integration of LLM reasoning with state-of-the-art geometric generative models. Code, data, and demos are available at https://smiles724.github.io/r1/.

cs.LG↗

arXiv2026

RiboSphere: Learning Unified and Efficient Representations of RNA Structures

Accurate RNA structure modeling remains difficult because RNA backbones are highly flexible, non-canonical interactions are prevalent, and experimentally determined 3D structures are comparatively scarce. We introduce RiboSphere, a framework that learns discrete geometric representations of RNA by combining vector quantization with flow matching. Our design is motivated by the modular organization of RNA architecture: complex folds are composed from recurring structural motifs. RiboSphere uses a geometric transformer encoder trained using mean-centered coordinates and random rotation augmentation to produce geometry-aware features, which are discretized with finite scalar quantization (FSQ) into a finite vocabulary of latent codes. Conditioned on these discrete codes, a flow-matching decoder reconstructs atomic coordinates, enabling high-fidelity structure generation. We find that the learned code indices are enriched for specific RNA motifs, suggesting that the model captures motif-level compositional structure rather than acting as a purely compressive bottleneck. Across benchmarks, RiboSphere achieves strong performance in structure reconstruction (RMSD 1.25,{\AA}, TM-score 0.84), and its pretrained discrete representations transfer effectively to inverse folding and RNA--ligand binding prediction, with robust generalization in data-scarce regimes. Code is available at https://github.com/Zhangz312/RiboSphere.

cs.LG↗

arXiv2026

Bi-TEAM: A Unified Cross-Scale Representation Learning Framework for Chemically Modified Biomolecules

Representation learning for protein biochemical space faces a difficult trade-off: protein language models excel at capturing long-range biological semantics but often miss fine-grained chemical details. Conversely, chemical language models encode atomic information but lack broader sequence context. To address this, we introduce Bi-TEAM (Bi-gated Residual Space Modification), a general framework that injects localized chemical variation into global protein contexts. By ensuring robustness against perturbations such as non-canonical amino acids, post-translational modifications (PTMs), and topological constraints, Bi-TEAM uncovers functional chemical dependencies often missed by evolutionary baselines. Mechanistically, Bi-TEAM maps non-canonical residues to their natural counterparts and injects atomic-level data via a bi-gated residual fusion mechanism. Crucially, this process uses modification-aware prompts to ensure that local structural changes influence global functional representations without requiring alphabet expansion. We evaluated Bi-TEAM on ten datasets spanning chemically modified peptides, PTMs, and natural proteins. The model consistently outperformed state-of-the-art baselines, achieving up to a 66 percent improvement in Matthews correlation coefficient (MCC) on scaffold-similarity splits and a 350 percent increase in hemolysis prediction accuracy. Furthermore, when deployed as an oracle for generative modeling, Bi-TEAM nearly quadrupled the success rate for designing cell-penetrating cyclic peptides. By unifying biological semantics with chemical precision, Bi-TEAM provides a versatile foundation for machine learning driven exploration of peptide and protein biochemical space.

q-bio.BM↗

arXiv2025

CONFIDE: Hallucination Assessment for Reliable Biomolecular Structure Prediction and Design

Reliable evaluation of protein structure predictions remains challenging, as metrics like pLDDT capture energetic stability but often miss subtle errors such as atomic clashes or conformational traps reflecting topological frustration within the protein folding energy landscape. We present CODE (Chain of Diffusion Embeddings), a self evaluating metric empirically found to quantify topological frustration directly from the latent diffusion embeddings of the AlphaFold3 series of structure predictors in a fully unsupervised manner. Integrating this with pLDDT, we propose CONFIDE, a unified evaluation framework that combines energetic and topological perspectives to improve the reliability of AlphaFold3 and related models. CODE strongly correlates with protein folding rates driven by topological frustration, achieving a correlation of 0.82 compared to pLDDT's 0.33 (a relative improvement of 148\%). CONFIDE significantly enhances the reliability of quality evaluation in molecular glue structure prediction benchmarks, achieving a Spearman correlation of 0.73 with RMSD, compared to pLDDT's correlation of 0.42, a relative improvement of 73.8\%. Beyond quality assessment, our approach applies to diverse drug design tasks, including all-atom binder design, enzymatic active site mapping, mutation induced binding affinity prediction, nucleic acid aptamer screening, and flexible protein modeling. By combining data driven embeddings with theoretical insight, CODE and CONFIDE outperform existing metrics across a wide range of biomolecular systems, offering robust and versatile tools to refine structure predictions, advance structural biology, and accelerate drug discovery.

q-bio.BM↗

arXiv2025

Lost in Tokenization: Context as the Key to Unlocking Biomolecular Understanding in Scientific LLMs

Scientific Large Language Models (Sci-LLMs) have emerged as a promising frontier for accelerating biological discovery. However, these models face a fundamental challenge when processing raw biomolecular sequences: the tokenization dilemma. Whether treating sequences as a specialized language, risking the loss of functional motif information, or as a separate modality, introducing formidable alignment challenges, current strategies fundamentally limit their reasoning capacity. We challenge this sequence-centric paradigm by positing that a more effective strategy is to provide Sci-LLMs with high-level structured context derived from established bioinformatics tools, thereby bypassing the need to interpret low-level noisy sequence data directly. Through a systematic comparison of leading Sci-LLMs on biological reasoning tasks, we tested three input modes: sequence-only, context-only, and a combination of both. Our findings are striking: the context-only approach consistently and substantially outperforms all other modes. Even more revealing, the inclusion of the raw sequence alongside its high-level context consistently degrades performance, indicating that raw sequences act as informational noise, even for models with specialized tokenization schemes. These results suggest that the primary strength of existing Sci-LLMs lies not in their nascent ability to interpret biomolecular syntax from scratch, but in their profound capacity for reasoning over structured, human-readable knowledge. Therefore, we argue for reframing Sci-LLMs not as sequence decoders, but as powerful reasoning engines over expert knowledge. This work lays the foundation for a new class of hybrid scientific AI agents, repositioning the developmental focus from direct sequence interpretation towards high-level knowledge synthesis. The code is available at https://github.com/opendatalab-raiser/CoKE.

cs.AI↗

arXiv2025

Learning the PTM Code through a Coarse-to-Fine, Mechanism-Aware Framework

Post-translational modifications (PTMs) form a combinatorial "code" that regulates protein function, yet deciphering this code - linking modified sites to their catalytic enzymes - remains a central unsolved problem in understanding cellular signaling and disease. We introduce COMPASS-PTM, a mechanism-aware, coarse-to-fine learning framework that unifies residue-level PTM profiling with enzyme-substrate assignment. COMPASS-PTM integrates evolutionary representations from protein language models with physicochemical priors and a crosstalk-aware prompting mechanism that explicitly models inter-PTM dependencies. This design allows the model to learn biologically coherent patterns of cooperative and antagonistic modifications while addressing the dual long-tail distribution of PTM data. Across multiple proteome-scale benchmarks, COMPASS-PTM establishes new state-of-the-art performance, including a 122% relative F1 improvement in multi-label site prediction and a 54% gain in zero-shot enzyme assignment. Beyond accuracy, the model demonstrates interpretable generalization, recovering canonical kinase motifs and predicting disease-associated PTM rewiring caused by missense variants. By bridging statistical learning with biochemical mechanism, COMPASS-PTM unifies site-level and enzyme-level prediction into a single framework that learns the grammar underlying protein regulation and signaling.

cs.CE↗

arXiv2025

ODesign: A World Model for Biomolecular Interaction Design

Biomolecular interactions underpin almost all biological processes, and their rational design is central to programming new biological functions. Generative AI models have emerged as powerful tools for molecular design, yet most remain specialized for individual molecular types and lack fine-grained control over interaction details. Here we present ODesign, an all-atom generative world model for all-to-all biomolecular interaction design. ODesign allows scientists to specify epitopes on arbitrary targets and generate diverse classes of binding partners with fine-grained control. Across entity-, token-, and atom-level benchmarks in the protein modality, ODesign demonstrates superior controllability and performance to modality-specific baselines. Extending beyond proteins, it generalizes to nucleic acid and small-molecule design, enabling interaction types such as protein-binding RNA/DNA and RNA/DNA-binding ligands that were previously inaccessible. By unifying multimodal biomolecular interactions within a single generative framework, ODesign moves toward a general-purpose molecular world model capable of programmable design. ODesign is available at https://odesign.lglab.ac.cn ,

q-bio.BM↗

arXiv2025

RiboPO: Preference Optimization for Structure- and Stability-Aware RNA Design

Designing RNA sequences that reliably adopt specified three-dimensional structures while maintaining thermodynamic stability remains challenging for synthetic biology and therapeutics. Current inverse folding approaches optimize for sequence recovery or single structural metrics, failing to simultaneously ensure global geometry, local accuracy, and ensemble stability-three interdependent requirements for functional RNA design. This gap becomes critical when designed sequences encounter dynamic biological environments. We introduce RiboPO, a Ribonucleic acid Preference Optimization framework that addresses this multi-objective challenge through reinforcement learning from physical feedback (RLPF). RiboPO fine-tunes gRNAde by constructing preference pairs from composite physical criteria that couple global 3D fidelity and thermodynamic stability. Preferences are formed using structural gates, PLDDT geometry assessments, and thermostability proxies with variability-aware margins, and the policy is updated with Direct Preference Optimization (DPO). On RNA inverse folding benchmarks, RiboPO demonstrates a superior balance of structural accuracy and stability. Compared to the best non-overlap baselines, our multi-round model improves Minimum Free Energy (MFE) by 12.3% and increases secondary-structure self-consistency (EternaFold scMCC) by 20%, while maintaining competitive 3D quality and high sequence diversity. In sampling efficiency, RiboPO achieves 11% higher pass@64 than the gRNAde base under the conjunction of multiple requirements. A multi-round variant with preference-pair reconstruction delivers additional gains on unseen RNA structures. These results establish RLPF as an effective paradigm for structure-accurate and ensemble-robust RNA design, providing a foundation for extending to complex biological objectives.

q-bio.BM↗

arXiv2025

From Supervision to Exploration: What Does Protein Language Model Learn During Reinforcement Learning?

Protein language models (PLMs) have advanced computational protein science through large-scale pretraining and scalable architectures. In parallel, reinforcement learning (RL) has broadened exploration and enabled precise multi-objective optimization in protein design. Yet whether RL can push PLMs beyond their pretraining priors to uncover latent sequence-structure-function rules remains unclear. We address this by pairing RL with PLMs across four domains: antimicrobial peptide design, kinase variant optimization, antibody engineering, and inverse folding. Using diverse RL algorithms and model classes, we ask if RL improves sampling efficiency and, more importantly, if it reveals capabilities not captured by supervised learning. Across benchmarks, RL consistently boosts success rates and sample efficiency. Performance follows a three-factor interaction: task headroom, reward fidelity, and policy capacity jointly determine gains. When rewards are accurate and informative, policies have sufficient capacity, and tasks leave room beyond supervised baselines, improvements scale; when rewards are noisy or capacity is constrained, gains saturate despite exploration. This view yields practical guidance for RL in protein design: prioritize reward modeling and calibration before scaling policy size, match algorithm and regularization strength to task difficulty, and allocate capacity where marginal gains are largest. Implementation is available at https://github.com/chq1155/RL-PLM.

cs.LG↗

arXiv2025

Position: The Hidden Costs and Measurement Gaps of Reinforcement Learning with Verifiable Rewards

Reinforcement learning with verifiable rewards (RLVR) is a practical, scalable way to improve large language models on math, code, and other structured tasks. However, we argue that many headline RLVR gains are not yet well validated because reports often conflate policy improvement with three confounds: (i) budget mismatch between RLVR and baseline evaluations, (ii) attempt inflation and calibration drift that convert abstentions into confident answers, and (iii) benchmark data contamination. Using budget-matched reproductions and partial-prompt contamination probes, we find that several widely cited gaps shrink substantially or disappear once budgets, prompts, and dataset versions are matched and contaminated sets are treated as memorization probes rather than evidence of reasoning. This does not mean that RLVR is ineffective, but it implies that current measurements often overstate capability gains and obscure reliability costs. We therefore propose a compact, tax-aware minimum standard for RLVR training and evaluation: budget-matched saturation curves with variance, calibration, and abstention tracking, a judge-robustness stress test when LLM judges are used, and an explicit contamination screen. With these controls, RLVR remains effective and deployable in verifiable domains, but reasoning gains should be treated as provisional without them.

cs.LG↗

arXiv2025

A deep reinforcement learning platform for antibiotic discovery

Antimicrobial resistance (AMR) is projected to cause up to 10 million deaths annually by 2050, underscoring the urgent need for new antibiotics. Here we present ApexAmphion, a deep-learning framework for de novo design of antibiotics that couples a 6.4-billion-parameter protein language model with reinforcement learning. The model is first fine-tuned on curated peptide data to capture antimicrobial sequence regularities, then optimised with proximal policy optimization against a composite reward that combines predictions from a learned minimum inhibitory concentration (MIC) classifier with differentiable physicochemical objectives. In vitro evaluation of 100 designed peptides showed low MIC values (nanomolar range in some cases) for all candidates (100% hit rate). Moreover, 99 our of 100 compounds exhibited broad-spectrum antimicrobial activity against at least two clinically relevant bacteria. The lead molecules killed bacteria primarily by potently targeting the cytoplasmic membrane. By unifying generation, scoring and multi-objective optimization with deep reinforcement learning in a single pipeline, our approach rapidly produces diverse, potent candidates, offering a scalable route to peptide antibiotics and a platform for iterative steering toward potency and developability within hours.

cs.LG↗