Journal of Chemical Information and Modeling 最新文献

{"title":"SFM-Net: Selective Fusion of Multiway Protein Feature Network for Predicting Binding Affinity Changes upon Mutations.","authors":"Chunting Liu, Sudong Cai, Tong Pan, Hiroyuki Ogata, Jiangning Song, Tatsuya Akutsu","doi":"10.1021/acs.jcim.5c00130","DOIUrl":"https://doi.org/10.1021/acs.jcim.5c00130","url":null,"abstract":"<p><p>Accurately predicting the effect of mutations on protein-protein interactions (PPIs) is essential for understanding the protein structure and function, as well as providing insights into disease-causing mechanisms. Many recent popular approaches based on the three-dimensional structure of proteins have been proposed to predict the changes in binding affinity caused by mutations, i.e. ΔΔ<i>G</i>. However, how to effectively use the structural information to comprehensively exploit complex interactions within proteins and integrate multisource features remains a significant challenge. In this study, we propose SFM-Net, a powerful deep learning model constructed with GNN-based multiway feature extractors and a new context-aware selective fusion module that jointly leverages the sequence, structural, and evolutionary information. Such design enables SFM-Net to effectively and selectively use features from different sources to facilitate binding affinity change prediction. Benchmarking experiments and targeted ablation studies illustrate the effectiveness and robustness of our method for improving the binding affinity change prediction.</p>","PeriodicalId":44,"journal":{"name":"Journal of Chemical Information and Modeling ","volume":" ","pages":""},"PeriodicalIF":5.6,"publicationDate":"2025-03-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143661732","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"化学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Predicting Fluorescence Emission Wavelengths and Quantum Yields via Machine Learning.","authors":"Rubens C Souza, Julio C Duarte, Ronaldo R Goldschmidt, Itamar Borges","doi":"10.1021/acs.jcim.4c02403","DOIUrl":"https://doi.org/10.1021/acs.jcim.4c02403","url":null,"abstract":"<p><p>The search for functional fluorescent organic materials can significantly benefit from the rapid and accurate predictions of photophysical properties. However, screening large numbers of potential fluorophore molecules in different solvents faces limitations of quantum mechanical calculations and experimental measurements. In this work, we develop machine learning (ML) algorithms for predicting the fluorescence of a molecule, focusing on two target properties: emission wavelengths (WLs) and quantum yields (QYs). For this purpose, we employ the Deep4Chem database which contains the optical properties of 20,236 combinations of 7,016 chromophores in 365 different solvents. Several chemical descriptors, or features, were selected as inputs for each model, and each molecule was characterized by its SMILES fingerprint. The Shapley additive explanations (SHAP) technique was used to rationalize the results, showing that the most impactful properties are chromophore-related, as expected from chemical intuition. For the best-performing model, the Random Forest, our results for the test set show a root-mean-square error (RMSE) of 28.8 nm (0.15 eV) for WLs and 0.19 for QYs. The developed ML models were used to predict, thus completing, the missing results for the WL and QY target properties in the original Deep4Chem database, resulting in two new databases: one for each property. Testing our ML models for each target property in molecules not included in the original Deep4Chem database gave good results.</p>","PeriodicalId":44,"journal":{"name":"Journal of Chemical Information and Modeling ","volume":" ","pages":""},"PeriodicalIF":5.6,"publicationDate":"2025-03-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143668482","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"化学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Leveraging-Induced Polarization for Drug Discovery: Efficient IC50 Prediction Using Minimal Features.","authors":"Ashraf Mohamed, Bernard R Brooks, Muhamed Amin","doi":"10.1021/acs.jcim.5c00076","DOIUrl":"https://doi.org/10.1021/acs.jcim.5c00076","url":null,"abstract":"<p><p>Here, we use the frequency of the atomic hybridizations (s, sp, sp², and sp³) of each atom type (H, C, N, O, S, etc.) within a molecule to predict the IC50s of drug-like molecules, focusing on compounds targeting the Thrombin, Estrogen Receptor alpha, and Phosphodiesterase 5A proteins. The Neural Network and Random Forest models yield high correlation coefficients (<i>R</i>²) and low mean square error (MSE) using only 19 features. The atomic hybridizations have been used previously to calculate the molecular polarizability using a simple empirical model (Miller et al. <i>JACS</i> <b>1979</b>). We show that the atomic hybridizations may also be used to accurately predict the molecular polarizabilities of these molecules. The results show the importance of the induced polarization in protein-ligand binding. Furthermore, the variation in <i>R</i>² and MSE for the different target proteins indicates that the contribution of the induced polarization to the binding energies is different for different target proteins.</p>","PeriodicalId":44,"journal":{"name":"Journal of Chemical Information and Modeling ","volume":" ","pages":""},"PeriodicalIF":5.6,"publicationDate":"2025-03-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143668481","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"化学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Unsupervised Machine Learning-Based Image Recognition of Raw Infrared Spectra: Toward Chemist-like Chemical Structural Classification and Beyond Numerical Data.","authors":"Kentarou Fuku, Takefumi Yoshida","doi":"10.1021/acs.jcim.4c01644","DOIUrl":"https://doi.org/10.1021/acs.jcim.4c01644","url":null,"abstract":"<p><p>Recent advances in artificial intelligence have significantly improved spectral data analysis. In this study, we used unsupervised machine learning to classify chemical compounds based on infrared (IR) spectral images, without relying on prior chemical knowledge. The potential of machine learning for chemical classification was demonstrated by extracting IR spectral images from the Spectral Database for Organic Compounds and converting them into 208,620-dimensional vector data. Hierarchical clustering of 230 compounds revealed distinct main clusters (<b>A</b>-<b>G</b>), each with specific subclusters exhibiting higher intracluster similarities. Despite the challenges, including sensitivity to spectral deviations and difficulty of distinguishing delicate chemical structures in spectra with low transparency in the fingerprint area, the proposed image recognition approach exhibits good potential. Both principal component analysis and k-means clustering produced similar results. Furthermore, the method demonstrated high robustness to noise. The Tanimoto coefficient was used to evaluate the molecular similarity, providing valuable insights. However, some results deviated from chemists' intuitions. The study also highlighted that the scaling composition formulas and molecular weights did not affect the classification results because high-dimensional features dominated the process. A comparison of the clustering results obtained from molecular fingerprints, using the adjusted Rand index as a metric, indicated that the image data provided better classification performance than numerical data of the same resolution. Overall, this study demonstrates the feasibility of using machine learning with IR spectral image data for chemical classification and offers a novel perspective that complements traditional methods, although the classifications may not always align with chemists' intuitions. This approach has broader implications for fields such as drug discovery, materials science, and automated spectral analysis, where handling large, raw spectral data sets is essential.</p>","PeriodicalId":44,"journal":{"name":"Journal of Chemical Information and Modeling ","volume":" ","pages":""},"PeriodicalIF":5.6,"publicationDate":"2025-03-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143661745","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"化学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Regulatory Interactions between APOBEC3B N- and C-Terminal Domains.","authors":"Mac Kevin E Braza, Özlem Demir, Surl-Hee Ahn, Clare K Morris, Carla Calvó-Tusell, Kelly L McGuire, Bárbara de la Peña Avalos, Michael A Carpenter, Yanjun Chen, Lorenzo Casalino, Hideki Aihara, Mark A Herzik, Reuben S Harris, Rommie E Amaro","doi":"10.1021/acs.jcim.4c02272","DOIUrl":"10.1021/acs.jcim.4c02272","url":null,"abstract":"<p><p>APOBEC3B (A3B) is implicated in DNA mutations that facilitate tumor evolution. Although structures of its individual N- and C-terminal domains (NTD and CTD) have been resolved through X-ray crystallography, the full-length A3B (fl-A3B) structure remains elusive, limiting our understanding of its dynamics and mechanisms. In particular, the APOBEC3B C-terminal domain (A3Bctd) is frequently closed in models and structures. In this study, we built several new models of fl-A3B using integrative structural biology methods and selected a top model for further dynamical investigation. We compared the dynamics of the truncated (A3Bctd) to that of the fl-A3B via conventional and Gaussian accelerated molecular dynamics (MD) simulations. Subsequently, we employed weighted ensemble methods to explore the fl-A3B active site opening mechanism, finding that interactions at the NTD-CTD interface enhance the opening frequency of the fl-A3B active site. Our findings shed light on the structural dynamics and potential druggability of fl-A3B, including observations regarding both the active and allosteric sites, which may offer new avenues for therapeutic intervention in cancer.</p>","PeriodicalId":44,"journal":{"name":"Journal of Chemical Information and Modeling ","volume":" ","pages":""},"PeriodicalIF":5.6,"publicationDate":"2025-03-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143655484","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"化学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Identification and Experimental Validation of NETosis-Mediated Abdominal Aortic Aneurysm Gene Signature Using Multi-omics, Machine Learning, and Mendelian Randomization.","authors":"Chengsong Wu, Yuanyuan Ren, Yang Li, Yue Cui, Liyao Zhang, Pan Zhang, Xuejiao Zhang, Shangguang Kan, Chan Zhang, Yuyan Xiong","doi":"10.1021/acs.jcim.4c02318","DOIUrl":"https://doi.org/10.1021/acs.jcim.4c02318","url":null,"abstract":"<p><p>Abdominal aortic aneurysm (AAA) is a life-threatening disorder with limited therapeutic options. Neutrophil extracellular traps (NETs) are formed by a process known as \"NETosis\" that has been implicated in AAA pathogenesis, yet the roles and prognostic significance of NET-related genes in AAA remain poorly understood. This study aimed to identify key AAA- and NET-related genes (AAA-NETs-RGs), elucidate their potential mechanisms in contributing to AAA, and explore potential therapeutic compounds for AAA therapy. Through bioinformatics analysis of multiomics and machine learning, we identified six AAA-NETs-RGs: DUSP26, FCN1, MTHFD2, GPRC5C, SEMA4A, and CCR7, which exhibited strong diagnostic potential for predicting AAA progression, were significantly enriched in pathways related to cytokine-cytokine receptor interaction and chemokine signaling. Immune infiltration analysis revealed a causal association between AAA-NETs-RGs and immune cell infiltration. Cell-cell communication analysis indicated that AAA-NETs-RGs predominantly function in smooth muscle cells, B cells, T cells, and NK cells, primarily through cytokine and chemokine signaling. Gene profiling revealed that CCR7 and MTHFD2 exhibited the most significant upregulation in AAA patients compared to non-AAA controls, as well as in <i>in vitro</i> AAA models. Notably, genetic depletion of CCR7 and MTHFD2 strongly inhibited Ang II-induced phenotypic switching, functional impairment, and senescence in vascular smooth muscle cells (VSMCs). Based on AAA-NETs-RGs, molecular docking analysis combined with the Connectivity Map (CMap) database identified mirdametinib as a potential therapeutic agent for AAA. Mirdametinib effectively alleviated Ang II-induced phenotypic switching, biological dysfunction, and senescence. These findings provide valuable insights into understanding the pathophysiology of AAA and highlight promising therapeutic strategies targeting AAA-NETs-RGs.</p>","PeriodicalId":44,"journal":{"name":"Journal of Chemical Information and Modeling ","volume":" ","pages":""},"PeriodicalIF":5.6,"publicationDate":"2025-03-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143655481","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"化学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Integrated Virtual Screening Approach Identifies New CYP19A1 Inhibitors.","authors":"Sijie Liu, Jie Wu, Ya Chen, Clemens Alexander Wolf, Matthias Bureik, Johannes Kirchmair, Mario Andrea Marchisio, Gerhard Wolber","doi":"10.1021/acs.jcim.5c00204","DOIUrl":"https://doi.org/10.1021/acs.jcim.5c00204","url":null,"abstract":"<p><p>The human cytochrome P450 19A1 (CYP19A1, aromatase) is a heme-containing protein catalyzing the final steps of the biosynthesis of the steroid hormone 17β-estradiol. It is a key target for the treatment of sex-hormone-related disorders due to its role in mediating the conversion of androgens to estrogens. Here, we report the development of a virtual screening workflow incorporating machine learning and structure-based modeling that has led to the discovery of new CYP19A1 inhibitors. The machine learning models were built on comprehensive CYP19A1 data sets extracted from the ChEMBL and PubChem Bioassay databases and subjected to thorough validation routines. Ten promising hits that resulted from the virtual screening campaign were selected for experimental testing in an enzymatic assay based on heterologous expression of human CYP19A1 in yeast. Among the seven structurally diverse compounds identified as new CYP19A1 inhibitors, compound <b>9</b>, a novel, noncovalent inhibitor based on coumarin and imidazole substructures, stood out by its high potency, with an IC₅₀ value of 271 ± 51 nM.</p>","PeriodicalId":44,"journal":{"name":"Journal of Chemical Information and Modeling ","volume":" ","pages":""},"PeriodicalIF":5.6,"publicationDate":"2025-03-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143661728","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"化学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"KUTE: Green-Kubo Uncertainty-Based Transport Coefficient Estimator.","authors":"Martín Otero-Lema, Raúl Lois-Cuns, Miguel A Boado, Hadrián Montes-Campos, Trinidad Méndez-Morales, Luis M Varela","doi":"10.1021/acs.jcim.4c02219","DOIUrl":"10.1021/acs.jcim.4c02219","url":null,"abstract":"<p><p>An algorithm for the calculation of transport properties from molecular dynamics simulations, kute, is introduced. The method estimates the integrals from the Green-Kubo theorem, taking into account the uncertainties of the correlation functions in order to eliminate arbitrary cutoffs or external parameters whose values could alter the result. In this contribution, the performance of kute is tested against other popular methods for the case of a protic ionic liquid for a variety of transport properties. It is found that kute achieves the same degree of accuracy as the equivalent formulation of the Einstein relations while performing better than other methods to calculate transport properties using Green-Kubo methods.</p>","PeriodicalId":44,"journal":{"name":"Journal of Chemical Information and Modeling ","volume":" ","pages":""},"PeriodicalIF":5.6,"publicationDate":"2025-03-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143655482","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"化学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"CPPCGM: A Highly Efficient Sequence-Based Tool for Simultaneously Identifying and Generating Cell-Penetrating Peptides.","authors":"Qiufen Chen, Yuewei Zhang, Jiali Gao, Jun Zhang","doi":"10.1021/acs.jcim.5c00199","DOIUrl":"https://doi.org/10.1021/acs.jcim.5c00199","url":null,"abstract":"<p><p>Cell-penetrating peptides (CPPs) are usually short oligopeptides with 5-30 amino acid residues. CPPs have been proven as important drug delivery vehicles into cells through different mechanisms, demonstrating their potential as therapeutic candidates. However, experimental screening and synthesis of CPPs could be time-consuming and expensive. Recently, numerous attempts have been made to develop computational methods as a cost-effective way for screening a number of potential CPP candidates. Despite significant advancements, current methods exhibit limited feature representation capabilities, thereby constraining the potential for further performance enhancements. In this study, we developed a deep learning framework called CPPCGM, which uses protein language models (PLMs) to identify and generate novel CPPs. There are two separate blocks in this framework: CPPClassifier and CPPGenerator. The former utilizes three pretrained models for simple voting, thereby accurately categorizing CPPs and non-CPPs. The latter, similar to a generative adversarial network, including a discriminator and a generator, generates peptides that are not present in the training data set. Our proposed CPPCGM has achieved remarkably high Matthews correlation coefficient scores of 0.876, 0.923, and 0.664 on three data sets based on the classification results. Compared with the state-of-the-art methods, the performance of our method is significantly improved. The results also demonstrated the generating potential of CPPCGM through qualitative and quantitative evaluation of the generated samples. Significantly, using PLM-based methods can optimize peptides for biochemical functions, benefiting drug delivery and biomedical applications. Materials related are publicly available at https://github.com/QiufenChen/CPPCGM.</p>","PeriodicalId":44,"journal":{"name":"Journal of Chemical Information and Modeling ","volume":" ","pages":""},"PeriodicalIF":5.6,"publicationDate":"2025-03-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143655474","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"化学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Assessing UFF and DFT-Tuned Force Fields for Predicting Experimental Isotherms of MOFs.","authors":"Yeongsu Cho, Jakob Teetz, Heather J Kulik","doi":"10.1021/acs.jcim.4c02044","DOIUrl":"https://doi.org/10.1021/acs.jcim.4c02044","url":null,"abstract":"<p><p>Metal-organic frameworks (MOFs) are promising materials for gas storage and separation applications due to their high tunability and porosity. The rational design of MOFs relies on accurate computational modeling, with grand canonical Monte Carlo (GCMC) simulations frequently employed to model gas uptake. However, GCMC predictions often deviate from experimental observations, limiting their utility in MOF screening. These discrepancies primarily arise from three factors: inaccuracies in the force field, neglect of atomic motions, and neglect of structural imperfections in MOFs. In this study, we systematically evaluate the impact of the first factor on the predictive accuracy of the GCMC simulations. We evaluate the widely used Universal Force Field (UFF) by comparing its predictions with experimental isotherms for four representative adsorbates, H₂, CO₂, C₂H₄, and C₂H₆, across 379 isotherms from 142 MOFs. The results show that UFF consistently overestimates the gas uptake in GCMC simulations. To isolate the contribution of force field inaccuracies to errors in GCMC, we developed a practical scheme for fitting force field parameters to DFT-calculated energies for a large set of MOFs. While the refined force field improves the accuracy of interatomic interaction energies, its reduction of repulsion, combined with UFF's tendency to overestimate gas uptake, ultimately amplifies the overestimation of experimental gas uptake meaurement. Our analysis suggests that improving the agreement of gas adsorption prediction with experiments requires addressing atomic motion and structural defects in MOFs alongside force field refinements.</p>","PeriodicalId":44,"journal":{"name":"Journal of Chemical Information and Modeling ","volume":" ","pages":""},"PeriodicalIF":5.6,"publicationDate":"2025-03-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143655472","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"化学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}