Journal of Cheminformatics最新文献

{"title":"A chemically-aware validation framework for benchmarking large language models in materials synthesis planning","authors":"Aobo Zhang","doi":"10.1186/s13321-026-01222-5","DOIUrl":"10.1186/s13321-026-01222-5","url":null,"abstract":"<p>We present a domain-tailored verification framework for evaluating the scientific quality of AI-generated synthesis protocols, moving beyond generic NLP benchmarks that fail to capture chemistry-specific requirements. Our approach combines two quantitative metrics: a framework score that assesses the logical coherence of the synthesis pathway, and a weighted detail score that measures the precision of reported experimental parameters.</p><p>This work establishes a benchmark for automated protocol generation, quantifies the gap between conceptual feasibility and parametric exactness in LLM outputs. We apply carefully curated dataset of SAC as a testbed to fine tune mainstream open source LLMs. The benchmark can be generalized to material synthesis protocols.</p>","PeriodicalId":617,"journal":{"name":"Journal of Cheminformatics","volume":"18 1","pages":""},"PeriodicalIF":7.9,"publicationDate":"2026-05-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13383416/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148013051","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"化学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Lignonaut: designing diverse combinatorial libraries for the exploration and annotation of lignin oligomer spaces","authors":"Mynta Norberg,&nbsp;Margareta Sandahl,&nbsp;Peter Spégel","doi":"10.1186/s13321-026-01202-9","DOIUrl":"10.1186/s13321-026-01202-9","url":null,"abstract":"<div><p>Lignins are a polymeric, renewable resource with remarkable structural diversity. Research is being conducted into the valorisation of lignins into value-added products. The absence of experimental libraries, in particular for process-modified oligomers, hampers analytical feedback to these valorisation efforts. Stochastic methods for generating libraries in-silico have been proposed, but were not designed for use with popular techniques such as high-resolution mass spectrometry. To resolve this, we developed Lignonaut, which is a toolkit for designing lignin libraries through virtual combinatorial synthesis. To ensure high interpretability we also developed new, diversity-oriented nomenclature for lignin oligomers, upon which an efficient SMILES translation algorithm could be built. Libraries of up to 10⁷ oligomers could be generated, in linear time, and at a rate of 10⁶ per minute.</p><p> <b>Scientific contribution</b></p><p>Lignonaut applies virtual combinatorial synthesis to exhaustively map lignin chemical spaces for e.g. high-resolution mass spectrometry. It is fast and can account for the high degree of isomerism in lignin oligomers, which were major limitations of previous stochastic approaches.</p></div>","PeriodicalId":617,"journal":{"name":"Journal of Cheminformatics","volume":"18 1","pages":""},"PeriodicalIF":5.7,"publicationDate":"2026-05-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://link.springer.com/content/pdf/10.1186/s13321-026-01202-9.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147985926","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"化学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"ExPO: an exposure-conditioned neural operator for L1000 signature prediction","authors":"Austin Spadaro,&nbsp;Alok Sharma,&nbsp;Iman Dehzangi","doi":"10.1186/s13321-026-01226-1","DOIUrl":"10.1186/s13321-026-01226-1","url":null,"abstract":"<div><p>Experimental profiling of drug–cell transcriptomic responses over dose and time is sparse and irregular, complicating discovery. We present ExPO, an exposure-conditioned neural operator that predicts full L1000 (978-gene) z-score signatures for a given compound–cell context as a continuous function of exposure. ExPO ifanstantiates a DeepONet, fusing ChemBERTa-2 molecular embeddings (with LoRA adaptation) and a trunk over sinusoidal Fourier features, so signatures can be evaluated at arbitrary dose–time pairs without regridding. Training is discovery-aligned, combining robust regression with a two-list listwise objective to optimize early-rank gene ordering, plus lightweight priors (Sobolev smoothness and an optional dose-monotonicity penalty) for pharmacologic plausibility. On a scaffold-held-out CMap/L1000 benchmark, ExPO improves over strong baselines in both accuracy and ranking (MAE 0.83 vs 0.89, Spearman 0.52 vs 0.48 vs DeepCE; NDCG@50↑/↓ 0.74/0.72 vs 0.71/0.69 for CIGER), tolerates withheld exposures (ΔMAE + 0.03 for interior interpolation; + 0.07 at edges, reduced by − 0.02 with the monotonicity prior), and transfers across cell lines (LCL-O MAE 0.90 vs 0.95; ρ 0.44 vs 0.40). Quantile heads with conformal calibration yield reliable uncertainty (test PICP 0.80/0.90), and filtering the least-confident 20% reduces MAE by ~ 18%. By modeling exposure as a field rather than discrete buckets, ExPO delivers accurate, rank-faithful, and calibrated signatures for in-silico exploration of dose–time surfaces. ExPO as a standalone predictor, along with its source code and benchmark dataset, is available at https://github.com/MLBC-lab/ExPO. </p></div>","PeriodicalId":617,"journal":{"name":"Journal of Cheminformatics","volume":"18 1","pages":""},"PeriodicalIF":7.9,"publicationDate":"2026-05-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://link.springer.com/content/pdf/10.1186/s13321-026-01226-1.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147986123","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"化学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Sequence-based drug-target binding site pre-training enables cryptic pocket detection and improves binding affinity and kinetics prediction","authors":"Shuo Zhang,&nbsp;Li Xie,&nbsp;Daniel Tiourine,&nbsp;Lei Xie","doi":"10.1186/s13321-026-01227-0","DOIUrl":"10.1186/s13321-026-01227-0","url":null,"abstract":"<div><p> Accurately characterizing protein-ligand binding, such as binding site, affinity and kinetics, is critical for accelerating drug discovery. However, many existing computational methods face key limitations, including insufficient integration of comprehensive databases, inadequate representation of protein structural dynamics, and incomplete modeling of microscale protein-ligand interactions. To address these challenges, we introduce ProMoNet, a sequence-based pre-training and fine-tuning framework to enhance the prediction of protein-ligand binding characteristics. ProMoNet leverages protein and molecular foundation models to expand data coverage and enhance diversity. It also introduces a pre-training strategy based on protein-ligand binding site prediction, which bridges protein- and ligand-level representations to support downstream prediction tasks involving protein-ligand complexes. Our pre-training module effectively models microscale protein-ligand interactions and captures the dynamic nature of proteins, including binding site crypticity, without relying on 3-dimensional structural inputs. Notably, this module surpasses or matches state-of-the-art structure-based methods in identifying exposed and cryptic binding sites while maintaining high efficiency. Our fine-tuning module then efficiently transfers the pre-trained knowledge to downstream tasks such as binding affinity and binding kinetics prediction, achieving superior performance. The combination of ProMoNet’s strong performance and demonstrated efficiency across multiple tasks highlights its potential for broad applications in drug discovery.</p><p><b>Scientific Contribution</b> We propose ProMoNet, a sequence-based pre-training and fine-tuning framework for protein-ligand binding characteristic prediction, where protein-ligand binding site prediction is introduced as a pre-training strategy to bridge independent protein- and ligand-level representations for downstream complex-level tasks. We design two dedicated modules, including a pre-training module that models microscale protein-ligand interactions and captures protein dynamics, as well as a fine-tuning module that efficiently integrates the pre-trained representations for downstream tasks. Even compared to structure-based methods, ProMoNet matches state-of-the-art performance in exposed and cryptic binding site identification and delivers superior results in binding affinity and kinetics prediction, making it a promising tool for drug discovery.</p></div>","PeriodicalId":617,"journal":{"name":"Journal of Cheminformatics","volume":"18 1","pages":""},"PeriodicalIF":7.9,"publicationDate":"2026-05-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://link.springer.com/content/pdf/10.1186/s13321-026-01227-0.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147980053","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"化学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"CataCon: a contrastive graph representation learning framework for catalyst prediction","authors":"Hua Shi,&nbsp;Yuzhe Wang,&nbsp;Shouzhen Song,&nbsp;Kefei Li,&nbsp;Wenjia Gao,&nbsp;Shuan Liu,&nbsp;Jianbo Qiao,&nbsp;Leyi Wei","doi":"10.1186/s13321-026-01212-7","DOIUrl":"10.1186/s13321-026-01212-7","url":null,"abstract":"<div><p>Catalysts accelerate chemical processes by lowering activation energy through alternative reaction pathways. However, existing prediction methods often struggle to capture intricate molecular information and overlook critical structural details such as atomic composition and bond topology. To address these challenges, we propose CataCon, a deep neural network framework grounded in contrastive learning. The model initially utilizes GraphSAGE to generate robust molecular graph embeddings for reactants, products, and candidate catalysts, thereby capturing their structural features comprehensively. Subsequently, a contrastive learning module constructs positive and negative sample pairs to align feature representations between reactant-product combinations and catalysts. This alignment enhances the model's capacity to identify potential interactions. Experimental results on public catalytic reaction datasets demonstrate that CataCon significantly outperforms baseline methods in catalyst classification tasks. Ablation studies further confirm the efficacy of combining graph representation learning with contrastive strategies for integrating multidimensional molecular information. Moreover, t-SNE visualization and interpretability analyses elucidate the underlying decision mechanisms of the model and provide novel perspectives for understanding the complex relationships among reaction components. <b>Scientific contribution</b> This study introduces CataCon, a novel contrastive graph representation learning framework designed to predict optimal catalysts for chemical reactions. This approach overcomes the limitations of existing methods that rely on simplistic catalyst labels by generating rich structural embeddings for all reaction components and aligning reaction and catalyst features through contrastive learning. By providing a powerful tool for the rational screening of catalyst candidates, CataCon achieves superior accuracy and has the potential to significantly accelerate materials discovery and optimize chemical synthesis.</p><h3>Graphical Abstract</h3>\u0000<div><figure><div><div><picture><source><img></source></picture></div></div></figure></div></div>","PeriodicalId":617,"journal":{"name":"Journal of Cheminformatics","volume":"18 1","pages":""},"PeriodicalIF":7.9,"publicationDate":"2026-05-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://link.springer.com/content/pdf/10.1186/s13321-026-01212-7.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147969442","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"化学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Novel molecular design via a scaffold-aware transformer with multi-scale attention mechanisms","authors":"Junyoung Park,&nbsp;Sunyong Yoo","doi":"10.1186/s13321-026-01221-6","DOIUrl":"10.1186/s13321-026-01221-6","url":null,"abstract":"<div><p>Recent advancements in artificial intelligence have demonstrated enhanced potential in accelerating drug discovery by exploring vast chemical spaces and predicting molecular properties. In particular, scaffold-constrained molecular generation has enabled the incorporation of structural constraints into the generation process. However, effectively integrating multi-scale structural information with activity-guided optimization remains challenging. To overcome these limitations, we propose a novel framework that integrates a transformer-based generative model and a graph attention network-based predictive model. The generative model produces molecules with desired structural characteristics by explicitly incorporating scaffold information, while the predictive model estimates the biological activity of the generated molecules. A supervised fine-tuning framework iteratively refines the generator through multi-stage tournament selection with experience memory. This framework guides the generator toward scaffold-consistent, high-affinity candidates while exploring novel chemical variations around a user-specified scaffold. Experimental results demonstrated that the proposed scaffold-aware transformer achieves competitive validity, uniqueness, and novelty, effectively generating novel compounds with high predicted binding affinity for biological targets. Meanwhile, an attention-based analysis extracted atom-level importance scores, highlighting the substructures that contribute to the predicted binding affinity and providing interpretable insights into structure–activity relationships. This study provides a practical and interpretable tool for scaffold-conditioned molecular generation. </p></div>","PeriodicalId":617,"journal":{"name":"Journal of Cheminformatics","volume":"18 1","pages":""},"PeriodicalIF":7.9,"publicationDate":"2026-05-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://link.springer.com/content/pdf/10.1186/s13321-026-01221-6.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147969435","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"化学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Revisiting ADMET prediction reliability under real-world challenges in the foundation model era","authors":"Donghai Zhao,&nbsp;Yuchen Zhu,&nbsp;Zhenxing Wu,&nbsp;Yue Wan,&nbsp;Xiaonan Liu,&nbsp;Shuaiyan Li,&nbsp;Hongxia Xu,&nbsp;Tingjun Hou,&nbsp;Chang-Yu Hsieh","doi":"10.1186/s13321-026-01217-2","DOIUrl":"10.1186/s13321-026-01217-2","url":null,"abstract":"&lt;div&gt;&lt;p&gt;As AI-driven drug discovery fully enters the era of foundation models, druggability prediction is undergoing a profound transformation from being experience-driven to “data-driven.” While numerous molecular foundation models with diverse representation capabilities have emerged, there remains a lack of systematic empirical mapping regarding their frontier boundaries and intrinsic limitations when faced with real-world drug discovery challenges. To this end, this study establishes a testing benchmark encompassing four major challenges—data scarcity and Out-of-Distribution (OOD) generalization, class imbalance, “beyond Rule of 5” (bRo5) chemical space generalization, and activity cliffs—to perform large-scale benchmarking on cutting-edge architectures such as molecular foundation models (KPGT, Uni-Mol) and tabular foundation models (TabPFNv2). Our analysis reveals the current frontiers and limitations of the molecular AI field: in few-shot and OOD scenarios, the tabular foundation model TabPFNv2 demonstrates superior generalization capabilities that often outperform those of the evaluated customized molecular foundation models; to address extreme class imbalance, ensemble strategies based on undersampling provide a robust mitigation approach with a reasonable trade-off between recall and precision; and within complex bRo5 chemical spaces, the pre-trained Graph Neural Network (GNN) KPGT achieves higher predictive accuracy compared to other tested architectures in predicting cyclic peptide permeability when provided with sufficient samples. However, all model categories consistently struggle to effectively identify activity cliffs, highlighting a persistent challenge for the AI architectures evaluated in this study to capture the phenomenon where minute structural variations lead to dramatic shifts in properties. To tackle these complex real-world challenges, this study confirms that ensembling high-performance models of different modalities is an effective strategy for improving predictive robustness. Furthermore, to facilitate the low-cost screening and construction of the high-quality base models needed for such ensembles, we validated that the molecular property landscape roughness index can serve as a lightweight diagnostic tool. Taken together, these insights provide practical guidance for the evaluation and selection of contemporary druggability prediction models.&lt;/p&gt;&lt;p&gt;&lt;b&gt;Scientific contribution&lt;/b&gt;&lt;/p&gt;&lt;p&gt;This study establishes a systematic benchmark framework for druggability prediction, simulating four key real-world drug discovery scenarios—including the underexplored generalization challenge in the bRo5 chemical space—and comprehensively evaluating molecular foundation models, the tabular foundation model TabPFNv2, GNNs, Automated Machine Learning (AutoML) framework, and classical machine learning approaches, thereby overcoming limitations of prior benchmarks in task diversity and model coverage. Through large-scale train–evaluati","PeriodicalId":617,"journal":{"name":"Journal of Cheminformatics","volume":"18 1","pages":""},"PeriodicalIF":7.9,"publicationDate":"2026-05-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://link.springer.com/content/pdf/10.1186/s13321-026-01217-2.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147969457","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"化学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"strainedSMILES2xyz: a workflow for reliable 3D structures of strained molecules from SMILES","authors":"Tori Demuth,&nbsp;Julian Schnizer,&nbsp;Dennis Svatunek","doi":"10.1186/s13321-026-01219-0","DOIUrl":"10.1186/s13321-026-01219-0","url":null,"abstract":"<div><p>Accurate 3D structure generation from SMILES is essential for data-driven chemistry but often fails for strained ring systems. We introduce <b>strainedSMILES2xyz</b>, a Python workflow that improves conformer generation by relaxing RDKit constraints, exploring stereoisomer variants, and correcting errors using force-field refinement. Benchmarking on strained and unstrained rings shows that it outperforms existing tools, generating correct geometries in nearly all cases. The workflow is available as a Python package and Jupyter notebook.</p><p><b>Scientific Contribution</b></p><p>This work identifies a critical gap in automated SMILES-to-3D structure generation for strained molecules. Many established tools, including the widely used RDKit, frequently fail or produce incorrect geometries for these systems. By explicitly targeting these failure modes, the proposed approach enables reliable 3D structure generation for chemically relevant strained molecules within fully automated workflows.</p></div>","PeriodicalId":617,"journal":{"name":"Journal of Cheminformatics","volume":"18 1","pages":""},"PeriodicalIF":7.9,"publicationDate":"2026-05-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://link.springer.com/content/pdf/10.1186/s13321-026-01219-0.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147959390","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"化学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Deep learning for assay nuisance compound detection using a gated co-attention graph embedding model (CAGE-Fusion)","authors":"Siddhant Rath,&nbsp;Saswati Panda,&nbsp;Steven J. Berthel,&nbsp;Kelsey Briggs,&nbsp;James C. Sacchettini","doi":"10.1186/s13321-026-01207-4","DOIUrl":"10.1186/s13321-026-01207-4","url":null,"abstract":"<div><p>In drug discovery, nuisance compounds are chemicals that interfere with biochemical or cell-based assays, often generating misleading signals unrelated to true activity. These molecules operate through diverse, context-dependent mechanisms, ranging from colloidal aggregation to covalent reactivity. They are often referred to as Pan-Assay Interference Compounds (PAINS), signal interferers, or frequent hitters, depending on their mode of action, and they lead to costly false positives and resource misallocation. They are difficult to distinguish from genuine pharmacological activity using rigid substructure filters. Although deep learning has emerged as a powerful alternative, most existing architectures process molecular graphs or SMILES sequences in isolation. We introduce CAGE-Fusion (Co-Attention Graph Embedding Fusion), a multimodal framework that employs a gated co-attention mechanism to enforce bidirectional information exchange between graph-based and sequence-based encoders. This architecture enables the iterative refinement of atom and token embeddings, yielding a chemically coherent representation that captures cross-modal dependencies. CAGE-Fusion demonstrates competitive predictive performance across several MoleculeNet benchmarks and achieves a macro-averaged PR-AUC of 0.73 and ROC-AUC of 0.94 on a comprehensive nuisance compound classification task. Beyond prediction, the model visualizes attention-weighted regions that contribute to a classification decision.</p><p><b>Scientific contribution</b></p><p>We present CAGE-Fusion, a gated co-attention architecture that enables bidirectional information exchange between graph-based encoders and SMILES representations to learn cross-modal molecular embeddings. We evaluate the framework on standardized molecular property benchmarks and apply it in a focused case study on assay nuisance compound detection across four nuisance categories: aggregation, luciferase inhibition, reactivity, and promiscuity.</p></div>","PeriodicalId":617,"journal":{"name":"Journal of Cheminformatics","volume":"18 1","pages":""},"PeriodicalIF":7.9,"publicationDate":"2026-05-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://link.springer.com/content/pdf/10.1186/s13321-026-01207-4.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147947447","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"化学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"PL-PatchSurfer3: improved structure-based virtual screening for structure variation using 3D Zernike descriptors","authors":"Woong-Hee Shin,&nbsp;Yuki Kagaya,&nbsp;Wonkyeong Jang,&nbsp;Daisuke Kihara","doi":"10.1186/s13321-026-01224-3","DOIUrl":"10.1186/s13321-026-01224-3","url":null,"abstract":"<div><p>Structure-based virtual screening (SBVS) is a widely used approach in in silico drug discovery, requiring a receptor structure or binding site to predict a ligand's binding pose and affinity. Consequently, the performance of SBVS depends on the protein’s conformation. The most common SBVS method is protein–ligand docking, which relies on physics-based, empirical, or knowledge-based scoring functions to estimate binding affinity. However, these methods are highly sensitive to structural variations and conformational changes of the receptor structure. The performance of SBVS often drops substantially when the apo form of the receptor used for screening is different from its holo form. To address this issue, we previously introduced an SBVS method called PL-PatchSurfer. This method uses a surface representation to describe a binding site of receptors and drug molecules, which is more robust to structural changes, offering more stable screening performance. The surface of the molecules is segmented into patches, and their shapes and physicochemical properties, such as the electrostatic potential, are represented by the 3D Zernike descriptor, a compact and rotationally invariant mathematical representation. Here, we introduce PL-PatchSurfer3, an improved version of PL-PatchSurfer, which incorporates two key enhancements: a refined definition of hydrogen bond complementarity and the integration of visibility, which captures curvature information of a patch. Our evaluation demonstrates that the new program outperforms its predecessor and other SBVS methods while maintaining its robustness against receptor structure variations.</p><p><b>Scientific contribution</b></p><p>This study presents PL-PatchSurfer3, an improved surface patch–based virtual screening method that incorporates molecular surface shape, physicochemical properties, hydrogen-bonding compatibility, and visibility, capturing local curvature using rotationally invariant 3D Zernike descriptors. PL-PatchSurfer3 demonstrates robust performance across holo, apo, modeled, and AlphaFold-predicted receptor structures and is superior to or competitive with conventional approaches and recent deep learning–based methods.</p></div>","PeriodicalId":617,"journal":{"name":"Journal of Cheminformatics","volume":"18 1","pages":""},"PeriodicalIF":7.9,"publicationDate":"2026-05-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://link.springer.com/content/pdf/10.1186/s13321-026-01224-3.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147939819","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"化学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}