Truong Son Hy

Proteins interact with a wide range of biomolecular partners, yet identifying the interaction sites that mediate these events remains a major challenge for understanding cellular function and guiding drug discovery. Existing computational methods are typically designed for specific interaction classes and often rely on task-specific features. Here, we introduce PRIME-ISP, a structure-based framework for predicting interaction sites across diverse biomolecular interaction types using a unified architecture. PRIME-ISP represents proteins at five organizational levels, including solvent-excluded surfaces, atoms, residues, secondary-structure elements, and whole proteins, and integrates information across these scales through coupled bottom-up and top-down propagation. This hierarchical representation jointly captures local geometry and global structural context, enabling a single architecture to address chemically and geometrically distinct binding problems without task-specific modifications. We evaluate PRIME-ISP on protein, peptide, DNA, RNA, small-molecule, metal, and carbohydrate interaction-site prediction benchmarks. Across these tasks, PRIME-ISP consistently outperforms a recent unified sequence-based predictor and achieves performance comparable to specialized methods while requiring no interaction-specific features. These results demonstrate that hierarchical structural representations provide a flexible foundation for interaction-site prediction across diverse biomolecular interactions. Source code and data are publicly available at https://github.com/HySonLab/PRIME-ISP .

Predicting the structural consequences of single-site mutations is fundamental to understanding protein function, disease mechanisms, and protein engineering. However, existing benchmarks for mutation-induced structural change prediction are often affected by duplicate mutation events, highly similar structural pairs, and experimental-state confounders, which can inflate reported performance. We introduce M uta S truct , a curated benchmark of experimentally observed single-mutation protein structure pairs designed for rigorous evaluation of mutation-induced structural changes. M uta S truct combines residue-level sequence validation, coordinate-level quality control, context-aware filtering, event-level deduplication, and leakage-aware data splitting to improve dataset reliability and evaluation rigor. We benchmark sequence-only structure prediction, physics-based refinement, and learning-based refinement methods together with a Copy-WT diagnostic control. Copy-WT remains highly competitive on backbone-based metrics, while method rankings change substantially across local and mutation-site measures. These results show that overall structural fidelity does not necessarily imply accurate modeling of mutation-associated structural change, underscoring the need for further methodological development beyond current paradigms. The benchmark dataset is currently undergoing further curation and expansion and will be released in a future journal-version update. The codebase, trained models, and evaluation pipeline are publicly available at https://github.com/HySonLab/MutaStruct .

Accurate identification of interactions between protein residues and ligand functional groups is critical for understanding molecular recognition and guiding rational drug design. Existing deep learning approaches for protein–ligand interpretability typically rely on three-dimensional structural input or distance-based contact labels, which limit both their applicability and biological relevance. Here, we present LINKER, the first sequence-based model to predict residue-functional group interactions according to biologically defined interaction types, using only a protein sequence and the SMILES representation of the ligand. LINKER is trained via structure-supervised interaction learning, in which interaction labels are derived from three-dimensional protein–ligand complexes through functional group-based motif extraction. By representing ligands as ensembles of functional groups, the model emphasizes chemically meaningful substructures rather than mere spatial proximity. Importantly, LINKER requires only sequence-level input at inference, enabling large-scale applications in contexts where structural data are unavailable. Extensive experiments demonstrate that LINKER consistently outperforms established baselines, highlighting the utility of functional group abstractions and structure-based supervision for interpretable protein–ligand interaction prediction. Our source code is publicly available at: https://github.com/HySonLab/LINKER/.

Background Mental health conditions such as depression, anxiety, and stress are commonly assessed using self-reported questionnaires and limited wearable physiological measures. However, reliance on subjective reporting, restricted sensor modalities such as heart rate variability and electrodermal activity, and small or homogeneous datasets may limit generalizability. We aimed to evaluate whether wearable optical sensing of microcirculation and tissue metabolism enables objective assessment of stress-related mental health states. Methods We conducted a prospective observational study including 132 adults aged 18 to 94 years (58% female) from 19 countries. Participants underwent repeated fingertip measurements using a non-invasive wearable device combining laser Doppler flowmetry and fluorescence spectroscopy to capture microvascular perfusion and metabolic signals. Frequency-domain features were extracted using wavelet analysis. Depression, anxiety, and stress levels were assessed using a standardized 21-item questionnaire. Multiple machine learning models were evaluated under subject-wise validation, and model interpretability was assessed using Shapley-based feature attribution. Results Here we show that ensemble-based models distinguish individuals with stress-related symptoms from those without with a receiver operating characteristic area under the curve of 0.72 and a precision-recall area under the curve of 0.89 under subject-wise validation. Microcirculatory variability and metabolic fluorescence features contribute substantially to prediction performance. Demographic variables, including sex, age, body mass index, and heart rate, are associated with increased stress-related risk. Conclusions Wearable optical sensing combined with interpretable machine learning provides physiological signatures associated with stress-related mental health conditions. This framework supports development of scalable and data-driven tools for objective mental health monitoring.

Molecular interactions often involve higher-order relationships that cannot be fully captured by traditional graph-based models limited to pairwise connections. Hypergraphs naturally extend graphs by enabling multi-way interactions, making them well-suited for modeling complex molecular systems. In this work, we introduce EquiHGNN, an equivariant hypergraph neural network framework that integrates symmetry-aware representations to improve molecular modeling. By enforcing the equivariance under relevant transformation groups, our approach preserves geometric and topological properties, leading to more robust and physically meaningful representations. We examine a range of equivariant architectures and demonstrate that integrating symmetry constraints leads to notable performance gains on large-scale molecular datasets. Experiments across small and large molecules indicate that while higher-order interactions provide marginal gains for small systems, they surpass 2D graphs on larger ones. Incorporating geometric features into these higher-order structures further enhances performance, underscoring the critical role of spatial information in molecular representation learning. Our source code is available at https://github.com/HySonLab/EquiHGNN/.

Multiresolution Matrix Factorization (MMF) is unusual amongst fast matrix factorization algorithms in that it does not make a low rank assumption. This makes MMF especially well suited to modeling certain types of graphs with complex multiscale or hierarchical strucutre. While MMF promises to yields a useful wavelet basis, finding the factorization itself is hard, and existing greedy methods tend to be brittle. In this paper, we propose a "learnable" version of MMF that carfully optimizes the factorization using metaheuristics, specifically evolutionary algorithms and directed evolution, along with Stiefel manifold optimization through backpropagating errors. We show that the resulting wavelet basis far outperforms prior MMF algorithms and gives comparable performance on standard learning tasks on graphs. Furthermore, we construct the wavelet neural networks (WNNs) learning graphs on the spectral domain with the wavelet basis produced by our MMF learning algorithm. Our wavelet networks are competitive against other state-of-the-art methods in molecular graphs classification and node classification on citation graphs. We release our implementation at https://github.com/HySonLab/LearnMMF

Industrial IoT predictive maintenance requires systems capable of real-time anomaly detection without sacrificing interpretability or demanding excessive computational resources. Traditional approaches rely on static, offline-trained models that cannot adapt to evolving operational conditions, while LLM-based monolithic systems demand prohibitive memory and latency, rendering them impractical for on-site edge deployment. We introduce SEMAS, a self-evolving hierarchical multi-agent system that distributes specialized agents across Edge, Fog, and Cloud computational tiers. Edge agents perform lightweight feature extraction and pre-filtering; Fog agents execute diversified ensemble detection with dynamic consensus voting; and Cloud agents continuously optimize system policies via Proximal Policy Optimization (PPO) while maintaining asynchronous, non-blocking inference. The framework incorporates LLM-based response generation for explainability and federated knowledge aggregation for adaptive policy distribution. This architecture enables resource-aware specialization without sacrificing real-time performance or model interpretability. Empirical evaluation on two industrial benchmarks (Boiler Emulator and Wind Turbine) demonstrates that SEMAS achieves superior anomaly detection performance with exceptional stability under adaptation, sustains prediction accuracy across evolving operational contexts, and delivers substantial latency improvements enabling genuine real-time deployment. Ablation studies confirm that PPO-driven policy evolution, consensus voting, and federated aggregation each contribute materially to system effectiveness. These findings indicate that resource-aware, self-evolving 1multi-agent coordination is essential for production-ready industrial IoT predictive maintenance under strict latency and explainability constraints.

Complex diseases, such as cancer and neurodegeneration, feature interconnected pathways, making single-target therapies ineffective due to pathway redundancy and compensatory mechanisms. Polypharmacy, which combines multiple drugs to target distinct proteins, addresses this but often leads to drug-drug interactions, cumulative toxicity, and complex pharmacokinetics. To overcome these challenges, we introduce EVOSYNTH, a modular framework for multi-target drug discovery that combines latent evolution and synthesis-aware prioritization to generate and prioritize candidates with high translational potential. Latent evolution navigates a chemically and functionally informed latent space to identify candidates with strong predicted affinity across multiple targets. Synthesis-aware prioritization evaluates both retrosynthetic feasibility and the trade-off between synthetic cost and therapeutic reward, enabling practical and efficient candidate selection. Applied to dual inhibition of JNK3 and GSK3β in Alzheimer’s disease and PI3K and PARP1 in ovarian cancer, EVOSYNTH consistently outperforms baseline generative models, achieving higher predicted affinities, improved scaffold diversity, and lower synthesis costs. These findings highlight EVOSYNTH’s ability to integrate target-driven generation with practical synthesizability, establishing a scalable framework for multi-target and polypharmacological drug discovery. Our source code and data to reproduce all experiments are publicly available on GitHub at: https://github.com/HySonLab/EvoSynth.