Digital discovery最新文献

{"title":"Breaking the digital-physical barrier in organic electronics materials discovery: application to finding anti-Kasha molecules.","authors":"Ömer H Omar, Daniele Padula, Kate A Morrison, Krzysztof Pawlak, Alexander Ciupa, Dávid Havasi, Alessandro Troisi","doi":"10.1039/d6dd00402d","DOIUrl":"10.1039/d6dd00402d","url":null,"abstract":"<p><p>Molecules which display two optical emission bands are very rare as they violate Kasha's rule, which states that fluorescence is normally observed only from the lowest energy excited state of a specific spin multiplicity (<i>e.g.</i> S₁) due to efficient internal conversion from higher states after excitation. Here, we develop a combined virtual and experimental screening workflow to find new anti-Kasha emitters among those with a large energy gap between S₁ and S₂; a condition that favours optical emission from the higher excited state because of a reduced internal conversion rate. 140 000 chemically diverse and commercially available molecules are filtered through a computational funnel based on their computed S₁-S₂ gap, transition intensities and structural rigidity before being ranked by cost and purity from up-to-date information from the chemical supply chain. The absorption and fluorescence spectra have been measured for 15 potential candidates that have not been previously explored for this property. Seven of them display anti-Kasha emission, five displaying single emission from S₂ and two displaying emission from multiple excited states (a particular case of dual emission). The lower energy emission of one of the dual emitters, however, originates from an aggregate state. The approach exemplifies the great potential of virtual screening approaches that take into account chemical availability and lead to an immediate validation of the predictions.</p>","PeriodicalId":72816,"journal":{"name":"Digital discovery","volume":" ","pages":"4463-4470"},"PeriodicalIF":7.1,"publicationDate":"2026-09-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13592029/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149018900","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"TRPM8 virtual screening: the essential role of target-specific, inactive-enriched machine-learning scoring functions.","authors":"Nivya James, Pedro J Ballester","doi":"10.1039/d6dd00399k","DOIUrl":"10.1039/d6dd00399k","url":null,"abstract":"<p><p>The Transient Receptor Potential Melastatin 8 (TRPM8) ion channel is an emerging therapeutic target implicated in pain, inflammation and other disorders. While structure-based virtual screening (VS) is widely used in drug discovery, the effectiveness of generic and target-specific machine-learning (ML) scoring functions (SFs) for ion channels such as TRPM8 remains to be investigated. We established the first comprehensive structure-based VS benchmark for TRPM8. This evaluates classical docking SFs, generic ML SFs and target-specific ML models retrospectively across multiple TRPM8 protein conformations and docking protocols. The evaluation revealed that generic ML rescoring at most provided modest and conformation-dependent improvements over classical docking on this target. Target-specific ML models based only on protein-ligand interaction fingerprints improved early enrichment on test sets when regression algorithms were employed. However, performance remained sensitive to protein conformation choice and chemical dissimilarity. Systematic evaluation of feature representations showed that combining structure-derived interaction features with ligand-based descriptors enhanced enrichment under selected docking tool-protein conformation combinations. Critically, presenting the learning algorithm with many more negative training instances, <i>via</i> inactive-enriched training, strongly reduced false positives. This also resulted in markedly improved generalization to chemically dissimilar test sets for structure-based models. By contrast, ligand-only QSAR models collapsed under the same class-imbalanced regime. Under inactive-enriched training, PLEC-based support vector regression (SVR) models achieved robust and strong VS performance, highlighting the critical role of learning the vast diversity of inactive molecules better during model training in developing predictive target-specific ML SFs for TRPM8 VS.</p>","PeriodicalId":72816,"journal":{"name":"Digital discovery","volume":" ","pages":"4434-4449"},"PeriodicalIF":7.1,"publicationDate":"2026-09-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13586914/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149009453","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Molecular docking <i>versus</i> co-folding for protein-peptide pose prediction: effects of protein flexibility and training set overlap.","authors":"Mark Fonteyne, Nada Badr, Peter J K Kuppen, Alexander L Vahrmeijer, Gerard J P van Westen, Willem Jespers","doi":"10.1039/d6dd00242k","DOIUrl":"10.1039/d6dd00242k","url":null,"abstract":"<p><p>Synthetic peptides are increasingly important as therapeutic and diagnostic agents due to their high specificity, ease in synthesis, and suitability for targeting protein interfaces. However, accurate prediction of peptide binding poses remains a major challenge, particularly for conformationally flexible protein targets and <i>de novo</i> designed peptides. Deep learning-based co-folding methods, including AlphaFold-Multimer (AFM), AlphaFold3 (AF3), and Boltz-2, offer an alternative to conventional peptide docking by simultaneously predicting protein-peptide complex structures without requiring predefined binding conformations. In this study, we systematically benchmarked these three co-folding methods against two peptide-optimized molecular docking methods, AutoDock-CrankPep (ADCP) and HADDOCK, using the PepPro dataset. This established dataset contains conformationally flexible protein-peptide complexes that are represented in the training sets of all co-folding methods. To assess the generalization performance of the co-folding methods, an independent test set was constructed classified by its level of redundancy with the training sets. The benchmark results showed that co-folding methods outperformed docking approaches in pose prediction accuracy and ranking performance when related complexes were represented in the training set. In the absence of such representation, however, HADDOCK outperformed all evaluated co-folding methods. Boltz-2 achieved the highest accuracy on the PepPro dataset but showed reduced generalization on non-redundant complexes, indicating strong dependence on training-set coverage. In contrast, AFM and AF3 demonstrated more robust generalization and maintained consistent performance on independent protein-peptide complexes, highlighting improved suitability for <i>de novo</i> peptide binder design. Furthermore, extensive analysis of scoring metrics revealed that interface predicted template modelling (ipTM) provides strong early enrichment performance, while peptide-specific predicted local distance difference test (pLDDT) correlates closely with structural accuracy, suggesting complementary value for candidate prioritization. These findings demonstrate that robust peptide-binder prediction requires careful consideration of training-set redundancy, target conformational flexibility, and complementary confidence metrics when selecting computational methods for <i>de novo</i> candidate prioritization.</p>","PeriodicalId":72816,"journal":{"name":"Digital discovery","volume":" ","pages":"4471-4488"},"PeriodicalIF":7.1,"publicationDate":"2026-09-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13592151/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149018843","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"moTSart: accelerating automated transition state search with generative models in a low-data regime.","authors":"Leonard Galustian, Johannes Karwounopoulos, Tori Demuth, Jasper De Landsheere, Konstantin Mark, Maximilian P-P Kovar, Anton Zamyatin, Dennis Svatunek, Esther Heid","doi":"10.1039/d6dd00259e","DOIUrl":"10.1039/d6dd00259e","url":null,"abstract":"<p><p>Transition states (TSs) are first-order saddle points on the potential energy surface, and thus the highest-energy structure on the (multi-step) minimum-energy reaction pathway that corresponds to a chemical reaction, determining the rate at which it proceeds. Locating TSs and thus elucidating reaction mechanisms is a computationally demanding and expertise-driven task. Recent automation strategies leverage heuristic rules or deep learning models to directly arrive at TS geometries from textual SMILES representations. However, standalone TS predictors lack robustness, whereas heuristic-based approaches lack scalability, limiting their applicability to reliable high-throughput reaction discovery under distribution shifts. Here, we present a generative machine learning framework embedded within a fully automated TS search pipeline and demonstrate accelerated high-throughput screening of bioorthogonal click reactions as a representative case study. The automated framework enables scalable exploration of large reaction spaces and progressively learns, reducing the number of required quantum mechanical evaluations as additional reactions are processed. In our case study, generative refinement decreased the required TS optimization cycles by nearly 40% after training on only 544 reactions. By coupling generative modeling with physics-based validation in a scalable workflow, our work establishes a framework for data-driven reaction mechanism exploration and advances the development of autonomous computational discovery of chemical reactivity.</p>","PeriodicalId":72816,"journal":{"name":"Digital discovery","volume":" ","pages":"4499-4511"},"PeriodicalIF":7.1,"publicationDate":"2026-09-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13592236/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149018855","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Pathway representation <i>via</i> intrinsic structural medoids (PRISM): a structural mapping approach to clustering molecular pathways.","authors":"Jherome Brylle Woody Santos, Jeremy M G Leung, Lillian T Chong, Ramón Alain Miranda-Quintana","doi":"10.1039/d6dd00288a","DOIUrl":"10.1039/d6dd00288a","url":null,"abstract":"<p><p>We present Pathway Representation <i>via</i> Intrinsic Structural Medoids (PRISM), a state-aware framework for clustering pathways from molecular dynamics simulations of biomolecular transitions. In PRISM, each pathway is mapped to a small set of structural medoids obtained <i>via</i> a deterministic <i>k</i>-means clustering scheme. Pairwise pathway dissimilarities are computed using a weighted average Hausdorff distance between these representative sets, effectively capturing mean nearest-neighbor structural deviations while reducing sensitivity to outliers. Hierarchical agglomerative clustering of the resulting dissimilarity matrix defines pathway families. We evaluate PRISM across three biomolecular transitions of increasing complexity: alanine dipeptide C7_eq → C7_ax isomerization, adenylate kinase opening, and HIF-2α PAS-B ligand unbinding. PRISM consistently yields robust cluster assignments, with medoids faithfully representing distinct conformational states. By combining a state-based description with robust geometric dissimilarities, PRISM provides a scalable framework for organizing complex transition pathways.</p>","PeriodicalId":72816,"journal":{"name":"Digital discovery","volume":" ","pages":"4110-4120"},"PeriodicalIF":7.1,"publicationDate":"2026-09-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13580032/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148992780","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Learning the reaction coordinate: collective variables from physical intuition to generative models.","authors":"Radu A Talmazan, Cheng Giuseppe Chen, Chenyu Tang, Alberto Megías, Sergio Contreras Arredondo, Christophe Chipot","doi":"10.1039/d6dd00290k","DOIUrl":"10.1039/d6dd00290k","url":null,"abstract":"<p><p>Collective variables (CVs) are low-dimensional projections of molecular configuration space that serve a dual purpose: they provide mechanistic interpretability by distilling complex transformations into comprehensible reaction coordinates, and they underpin the majority of enhanced sampling methods by defining the directions along which free-energy barriers are overcome. The discovery of suitable CVs has evolved from reliance on chemical intuition, for instance by selecting distances, angles, or dihedrals by hand, the systematic linear approaches, such as principal component analysis and time-lagged independent component analysis, to nonlinear manifold-learning techniques including diffusion maps and variational autoencoders. Deep-learning methods have further expanded this landscape: committor-based neural networks approximate the optimal reaction coordinate from trajectory data, discriminant and variational models learn CVs tied to metastable-state separation or slow kinetics, and equivariant graph neural networks construct symmetry-preserving representations directly from atomic coordinates. In parallel, generative models-normalizing flows, diffusion-based samplers, and learned transfer operators-have begun to bypass explicit dimensionality reduction altogether, learning equilibrium distributions or dynamical propagators in the full configurational space. Yet, these models encode latent structure from which CVs can be extracted <i>a posteriori</i> for mechanistic interpretation. In this Perspective, we trace the arc of CV discovery from intuition-driven heuristics to modern data-driven and generative frameworks, critically assess the strengths and limitations of each class of methods, and outline how the convergence of machine-learned potentials, automated CV learning, generative sampling, and causal interpretability is giving rise to integrated workflows that will reshape predictive molecular simulation across biomolecular, catalytic, and materials systems.</p>","PeriodicalId":72816,"journal":{"name":"Digital discovery","volume":" ","pages":"3984-3999"},"PeriodicalIF":7.1,"publicationDate":"2026-08-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13523130/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148852090","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Evaluating 2D and 3D graph encoders for infrared-based organic structure verification.","authors":"J Benji Rowlands, Richard J Lewis, Peter Howe, Jonathan M Goodman","doi":"10.1039/d6dd00281a","DOIUrl":"10.1039/d6dd00281a","url":null,"abstract":"<p><p>Infrared (IR) spectrum prediction is typically judged by global spectral error, but the practical task in analytical chemistry is structure verification: given a spectrum, can predicted spectra rank the correct structure highly among many candidates? We train Phase-Aware Spectrum Generative Network (PASGeN) models to predict experimental IR spectra and evaluate their predictions using both spectral similarity and retrieval metrics. Using 13 568 experimental IR spectra in various phases, we train models with a 2D message-passing neural network (MPNN) and a 3D equivariant encoder (PaiNN), and fine-tune the 3D model on inexpensive force-field geometries to reduce inference cost. Across large-scale retrieval tests, the 3D encoder significantly outperforms the 2D baseline (73.9% <i>vs.</i> 62.9% Top-1 accuracy). On this benchmark, both learned models achieve higher retrieval accuracy than the B3LYP/6-31G(d) and GFN2-xTB protocols evaluated here, which use gas-phase harmonic spectra with simple broadening. Descriptor analysis suggests that ring-rich, geometrically complex molecules benefit most from the inclusion of 3D information. By evaluating IR prediction models on a structure verification task and demonstrating that lightweight 3D geometries provide a measurable advantage, this work establishes a practical and scalable route to IR-based structure confirmation that outperforms reference physics-based baselines at a fraction of the computational cost.</p>","PeriodicalId":72816,"journal":{"name":"Digital discovery","volume":"5 10","pages":"4180-4190"},"PeriodicalIF":7.1,"publicationDate":"2026-08-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13667908/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149198257","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Tackling real-world crystal structure prediction from powder X-ray diffraction data.","authors":"Frederik Lizak Johansen, Adam F Sapnik, Erik Bjørnager Dam, Raghavendra Selvan, Kirsten M Ø Jensen","doi":"10.1039/d6dd00088f","DOIUrl":"10.1039/d6dd00088f","url":null,"abstract":"<p><p>Crystal structure prediction (CSP) from powder diffraction data is a central challenge in materials chemistry. Machine learning (ML) models show promise, but most are trained on idealized simulated data, limiting reliability on real experiments. Here, we assess real-world behaviour using the previously published deCIFer model as an example of PXRD-conditioned generative CSP. deCIFer is an autoregressive transformer that conditions each step of structure generation on encoded PXRD data, guiding token-wise predictions of space group, lattice parameters, and atomic positions. Using controlled robustness tests, we quantify performance under realistic artefacts (noise, background, peak asymmetry, and Scherrer broadening) and introduce metrics for accuracy and predictive uncertainty. deCIFer adapts smoothly to signal distortions and improves over unconditioned baselines when diffraction features remain informative, while expressing appropriate uncertainty as the PXRD pattern becomes underdetermined. Experimental PXRD tests recover the known structures of Si and CeO₂ and expose the expected limitations for lower-symmetry Fe₂O₃ and nanocrystalline CeO₂. Overall, ML-based CSP is fundamentally limited by the information content of PXRD, but can accelerate expert workflows by rapidly generating chemically plausible candidates and quantifying uncertainty, making such models valuable human-in-the-loop tools for real-world structure determination.</p>","PeriodicalId":72816,"journal":{"name":"Digital discovery","volume":" ","pages":"3871-3882"},"PeriodicalIF":7.1,"publicationDate":"2026-08-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13470798/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148766341","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Bridging data and theory: hybrid modelling for pharmaceutical crystallisation process development.","authors":"Thomas Pickles, Christos Xiouras, Merve Öner, Cameron J Brown","doi":"10.1039/d6dd00342g","DOIUrl":"https://doi.org/10.1039/d6dd00342g","url":null,"abstract":"<p><p>Crystallisation remains a critical unit operation in pharmaceutical manufacturing, yet process development is often constrained by empirical approaches and data limitations. Mechanistic models such as population balance frameworks provide physical interpretability but require extensive parameterisation and struggle with complex, multiscale phenomena. Data-driven machine learning models offer strong predictive performance but lack transparency, limiting regulatory acceptance. Hybrid modelling, which integrates mechanistic theory with data-driven components, has emerged as a promising strategy to address these limitations. This review surveys the current modelling landscape and critically assesses recent case studies across batch cooling, antisolvent, and continuous crystallisation modes, covering population balance and machine learning combinations. Key open problems are identified, including the absence of agreed modelling and experimental design frameworks, data sparsity, poor scale-up transferability, unresolved regulatory requirements around uncertainty quantification and model. Practical technical recommendations and a proposed workflow are provided to guide future work, with the aim of moving the field beyond isolated case studies toward predictive, explainable, deployable and published crystallisation process development.</p>","PeriodicalId":72816,"journal":{"name":"Digital discovery","volume":"5 9","pages":"3449-3458"},"PeriodicalIF":7.1,"publicationDate":"2026-07-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13580607/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148992764","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Censoring chemical data to mitigate dual use risk.","authors":"Quintina Campbell, Jonathan Herington, Andrew D White","doi":"10.1039/d5dd00512d","DOIUrl":"10.1039/d5dd00512d","url":null,"abstract":"<p><p>Machine learning models have dual use potential, potentially serving both beneficial and malicious purposes. The development of open-source models in chemistry has specifically surfaced dual use concerns around toxicological data and chemical warfare agents. We discuss a chain risk framework identifying three misuse pathways and corresponding mitigation strategies: inference-level, model-level, and data-level. At the data level, we introduce a noising method to increase prediction error in specific desired regions (sensitive regions). Our results show that selective noise induces variance and attenuation bias, whereas simply omitting sensitive data fails to prevent extrapolation. These findings hold for both molecular feature multilayer perceptrons and graph neural networks. Thus, noising molecular structures represents a step toward enabling safer sharing of potential dual use molecular data.</p>","PeriodicalId":72816,"journal":{"name":"Digital discovery","volume":" ","pages":"3275-3289"},"PeriodicalIF":7.1,"publicationDate":"2026-07-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13334314/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148439129","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}