相关参考文献
注意:仅列出部分参考文献,下载原文获取全部文献信息。
Article
Chemistry, Medicinal
Maria Vittoria Togo et al.
Summary: This article presents a robust and reproducible eXplainable Artificial Intelligence (XAI) approach for predicting developmental toxicity, which is a challenging human-health endpoint in toxicology. The proposed framework compares favorably with state-of-the-art approaches in terms of accuracy, sensitivity, and specificity, providing a reliable support system for ensuring informativeness, uncertainty estimation, generalization, and transparency in developmental toxicity. The model utilizes specific molecular descriptors and structural alerts to distinguish toxic and nontoxic chemicals.
JOURNAL OF CHEMICAL INFORMATION AND MODELING
(2023)
Article
Chemistry, Medicinal
Raghad Aljarf et al.
Summary: Teratogenic drugs can cause severe fetal malformation and have a critical impact on fetal health, but the teratogenic risks of most approved drugs are unknown. In this study, researchers propose a novel predictive tool called embryoTox, which uses a graph-based signature representation of chemical structure to predict and classify molecules that are likely to be safe during pregnancy. The tool achieved high accuracy on cross-validation and blind tests, outperforming alternative approaches. The authors believe that embryoTox will provide a practical resource for screening libraries and identifying safe molecules for use during pregnancy. The tool is freely available online at https://biosig.lab.uq.edu.au/embryotox/.
JOURNAL OF CHEMICAL INFORMATION AND MODELING
(2023)
Article
Chemistry, Multidisciplinary
O. J. M. Bequignon et al.
Summary: With the rapid growth of publicly available ligand-protein bioactivity data, researchers face challenges in adapting and finding the right data for their needs. To address this, the Papyrus dataset has been constructed, comprising around 60 million data points from various sources. The aggregated data has been standardized and normalized for machine learning purposes, and examples of data filtering and structure-activity relationship analyses are provided.
JOURNAL OF CHEMINFORMATICS
(2023)
Article
Chemistry, Medicinal
David Buterez et al.
Summary: High-throughput screening is a key technique in drug discovery for identifying potential drug candidates in an automated and cost-effective way. However, existing machine learning models for this process ignore a significant amount of experimental data, hindering drug development progress. To address this issue, we introduce the MF-PCBA dataset, which includes two data modalities for each dataset, primary and confirmatory screening, providing comprehensive and accurate data for drug discovery.
JOURNAL OF CHEMICAL INFORMATION AND MODELING
(2023)
Article
Chemistry, Medicinal
Hou Yee Choo et al.
Summary: This paper proposes a fingerprint-enhanced graph attention network (FinGAT) model, which combines sequence-based 2D fingerprints and structure-based graph representation to improve the accuracy of antibiotic discovery. In the feature learning process, sequence information is transformed into a fingerprint vector, and structural information is encoded through a GAT module into another vector. These two vectors are concatenated and input into a multilayer perceptron (MLP) for antibiotic activity classification. Our FinGAT has been extensively tested and compared with existing models, and it has been found to outperform various state-of-the-art GNN models in antibiotic discovery.
JOURNAL OF CHEMICAL INFORMATION AND MODELING
(2023)
Article
Biochemistry & Molecular Biology
Sunghwan Kim et al.
Summary: PubChem, a popular chemical information resource, has undergone several changes and improvements. It has added data from over 120 sources and introduced new functionalities, including the integration of Google Patents data, creation of Cell Line and Taxonomy data collections, and updates to the bioassay data model.
NUCLEIC ACIDS RESEARCH
(2023)
Article
Biochemical Research Methods
Jinsong Shao et al.
Summary: In this study, a natural language processing technique was used to analyze compounds, accurately representing the relationship between the compound and its substructure through a pre-trained optimized word2vec model. The machine learning model based on the training results effectively predicted the inhibitory effect of compounds on HBV and liver toxicity. The importance of this research lies in providing a new perspective on compound property prediction for anti-HBV drugs, which can improve hepatitis B diagnosis and further develop human health.
BRIEFINGS IN BIOINFORMATICS
(2022)
Review
Chemistry, Physical
Shampa Raghunathan et al.
Summary: Machine learning methods have become increasingly common in molecular sciences, especially in predicting chemical properties and designing molecules and materials. The unique encoding representation of molecular entities plays a crucial role in accurately predicting properties using ML models. Hierarchy of representations and diverse applications are instrumental in advancing the field of ML accelerated computational modeling.
INTERNATIONAL JOURNAL OF QUANTUM CHEMISTRY
(2022)
Article
Chemistry, Multidisciplinary
Andreas Luttens et al.
Summary: Developing drug inhibitors targeting SARS-CoV-2 is crucial for saving lives and preparing for future outbreaks. Two virtual screening strategies were explored, resulting in the identification of compounds with inhibitory effects, including one compound with promising antiviral activity.
JOURNAL OF THE AMERICAN CHEMICAL SOCIETY
(2022)
Article
Multidisciplinary Sciences
Arman A. Sadybekov et al.
Summary: Structure-based virtual ligand screening using V-SYNTHES method is effective in rapidly identifying high-scoring compounds from a large chemical space, with experimental results demonstrating successful hit rates and potencies. This approach shows promise for lead discovery and is easily scalable for use with diverse docking algorithms.
Article
Biochemical Research Methods
Francesco Gentile et al.
Summary: With the rapid expansion of chemical libraries, more efficient virtual screening approaches are needed. The Deep Docking (DD) platform allows for accelerated structure-based virtual screening by docking only a subset of the library and synchronizing it with ligand-based predictions. This method enables the screening of large chemical libraries without requiring extraordinary computational resources.
Review
Biochemical Research Methods
Maria Virginia Sabando et al.
Summary: With the consolidation of deep learning in drug discovery, several novel algorithms for learning molecular representations have been proposed. However, comparing different molecular embeddings and traditional representations is not straightforward, hindering the process of choosing suitable representations for QSAR modeling. The study conducted experiments comparing different embedding techniques and found that the predictive performance using molecular embeddings did not significantly surpass that of traditional representations.
BRIEFINGS IN BIOINFORMATICS
(2022)
Review
Pharmacology & Pharmacy
Nikhil Pillai et al.
Summary: Machine learning has been widely used in the early stages of drug discovery, but its applications in pharmacokinetic/pharmacodynamic (PK/PD) field are still limited. Recent progress in ML has focused on predicting ADME properties of small molecules and PK of drug candidates, providing important insights into safety and efficacy.
DRUG DISCOVERY TODAY
(2022)
Article
Chemistry, Medicinal
Xudong Zhang et al.
Summary: In this study, a series of hERG blocking classification models were trained using MACCS and Morgan fingerprints, and a more accurate consensus model named HergSPred was generated. Additionally, an analysis of the contribution of each input fingerprint to the prediction output provided chemical insights into hERG inhibition, allowing visualization of warning substructures that may cause cardiotoxicity.
JOURNAL OF CHEMICAL INFORMATION AND MODELING
(2022)
Article
Chemistry, Medicinal
Stevan Aleksic et al.
Summary: Despite the widespread use of computational methods in drug discovery and development in the pharmaceutical industry, our study found that there is no significant impact on prediction from data volume, modeling algorithm, chemical representation and grouping, and temporal aspect relationships.
MOLECULAR INFORMATICS
(2022)
Article
Pharmacology & Pharmacy
Vishwesh Venkatraman et al.
Summary: The SARS-CoV2 pandemic emphasizes the importance of efficient drug identification methods. Virtual screening methods have the potential to evaluate billions of candidate molecules, expanding the search space and speeding up discovery. This article describes a new screening pipeline called drugsniffer, capable of rapidly exploring drug candidates from a library of billions of molecules.
FRONTIERS IN PHARMACOLOGY
(2022)
Article
Chemistry, Medicinal
Marcus Tullius Scotti et al.
Summary: We introduce MolPredictX, an innovative and freely accessible web interface for predicting the biological activity of query molecules. MolPredictX utilizes in-house QSAR models to provide qualitative predictions and quantitative probabilities for a variety of diseases-related bioactivities.
MOLECULAR INFORMATICS
(2022)
Article
Chemistry, Multidisciplinary
Ruihan Zhang et al.
Summary: This article presents a comprehensive online web platform for anti-inflammatory natural product research. The platform includes a database of tested natural products with anti-inflammatory activity, as well as two predictive tools for predicting anti-inflammatory activity and compound-target relationships.
JOURNAL OF CHEMINFORMATICS
(2022)
Article
Chemistry, Multidisciplinary
Almos Orosz et al.
Summary: The study compared different descriptor groups for ADME-Tox classification targets and found that the traditional 1D, 2D, and 3D descriptors outperformed other descriptor sets, especially when using the XGBoost algorithm.
FRONTIERS IN CHEMISTRY
(2022)
Editorial Material
Pharmacology & Pharmacy
Ingo Muegge et al.
EXPERT OPINION ON DRUG DISCOVERY
(2022)
Article
Chemistry, Medicinal
Saba Iftkha et al.
Summary: Toxicity is a major concern in drug design, and current computational methods for predicting toxicity have limitations. To address this, we propose a new web-based computational method, cardioToxCSM, which accurately predicts cardiac toxicity outcomes.
JOURNAL OF CHEMICAL INFORMATION AND MODELING
(2022)
Article
Biology
Joseph Adams et al.
Summary: This study developed machine learning models to predict small molecules as potential anti-Ebola virus compounds capable of inhibiting glycoprotein and VP40 activities. The random forest model showed the best performance with an accuracy of 89%, an F1 score of 0.9, and a high AUC score of 0.95.
COMPUTATIONAL BIOLOGY AND CHEMISTRY
(2022)
Article
Computer Science, Artificial Intelligence
Jerret Ross et al.
Summary: Researchers have used large language models to model molecules and obtained good results in property prediction. By training an efficient transformer model on unlabeled molecular data, better molecular embeddings were obtained compared to existing benchmarks. This suggests that large-scale molecular language models can capture sufficient chemical and structural information to predict various distinct molecular properties.
NATURE MACHINE INTELLIGENCE
(2022)
Review
Biochemical Research Methods
Natesh Singh et al.
Summary: The interaction between life sciences and advancing technology leads to continuous growth of chemical data, with virtual screening methods being popular in pharmaceutical research. Web-based tools assist scientists in conducting virtual screening experiments, contributing to the design of bioactive molecules and drug development teaching.
BRIEFINGS IN BIOINFORMATICS
(2021)
Article
Biochemical Research Methods
Gabriel H. S. Dreiman et al.
Summary: Iterative screening uses machine learning to select promising compounds, improving efficiency and reducing the number of compounds screened. Results show iterative screening can identify 70%-90% of active compounds without requiring significant computational resources.
Article
Biochemical Research Methods
Zhenxing Wu et al.
Summary: A study on learning QSAR models using various ML algorithms for 14 public datasets showed that rbf-SVM, rbf-GPR, XGBoost, and DNN generally perform better than other algorithms. SVM and XGBoost are recommended for regression learning on small datasets, while XGBoost is an excellent choice for large datasets. Ensemble models integrating multiple algorithms can improve prediction accuracy.
BRIEFINGS IN BIOINFORMATICS
(2021)
Article
Medicine, Research & Experimental
Thomas R. Lane et al.
Summary: Machine learning methods are being widely used in the pharmaceutical industry for drug discovery. Studies have applied various algorithms and modeling metrics to build models for individual targets or properties using data from public databases. The performance of the models was evaluated using a range of cross-validation metrics, with different methods showing comparable results.
MOLECULAR PHARMACEUTICS
(2021)
Article
Chemistry, Medicinal
Carlos H. M. Rodrigues et al.
Summary: A machine learning approach using graph-based representation accurately identifies inhibitors modulating protein-protein interactions, leading to development of models that predict active compounds efficiently.
JOURNAL OF CHEMICAL INFORMATION AND MODELING
(2021)
Article
Biochemical Research Methods
Brian J. Bender et al.
Summary: Structure-based docking screens of compound libraries are common in early drug and probe discovery. Best practices and control calculations are outlined to evaluate docking parameters prior to undertaking a large-scale prospective screen.
Article
Chemistry, Multidisciplinary
Nicolas Bosc et al.
Summary: The research team developed a consensus in silico model for identifying anti-malarial molecules and addressed the challenge of data integration by sharing QSAR models. They developed an open-source software platform and launched a website called MAIP for the wider community to freely use for predicting potential malaria inhibiting compounds.
JOURNAL OF CHEMINFORMATICS
(2021)
Article
Chemistry, Multidisciplinary
Annachiara Tinivella et al.
Summary: The development of selective inhibitors for human Carbonic Anhydrase isoforms IX and XII is essential for cancer drug research. Machine learning classification models trained on high confidence data from ChEMBL were used to predict ligand activity and selectivity profiles. The results outperformed classic models built with a priori activity thresholds.
JOURNAL OF CHEMINFORMATICS
(2021)
Article
Chemistry, Multidisciplinary
Vishwesh Venkatraman
Summary: This article examines the efficacy of fingerprint-based machine learning models for a large number of ADMET-related properties. For most properties, fingerprint-based random forest models show comparable or better performance than traditional 2D/3D molecular descriptors.
JOURNAL OF CHEMINFORMATICS
(2021)
Article
Biochemical Research Methods
Qing Ye et al.
Summary: This study developed classification models using machine learning algorithms to distinguish Mtb inhibitors from noninhibitors, with the XGBoost model showing the best prediction performance. Two consensus strategies were employed to integrate predictions from multiple models, resulting in the best predictions. The association between important descriptors and bioactivities of molecules was interpreted, and an online tool called ChemTB was developed for detecting potential Mtb inhibitors.
BRIEFINGS IN BIOINFORMATICS
(2021)
Article
Biochemical Research Methods
Zhenxing Wu et al.
Summary: A new graph neural network model combining molecular graphs and molecular descriptors for drug discovery predictions is proposed, showing high efficiency and strong anti-noise capability.
BRIEFINGS IN BIOINFORMATICS
(2021)
Article
Pharmacology & Pharmacy
Shuyun He et al.
Summary: This study collected diverse datasets of active and inactive compounds for breast cancer cell lines and developed multiple predictive models using conventional and deep learning algorithms. The best model for each breast cancer cell subtype showed high predictive accuracy, with AUC values ranging from 0.689 to 0.993. Important structural fragments related to breast cancer cell inhibition were identified, and an online webserver and local version software were developed for predicting potential inhibitory activity against breast cancer cells.
FRONTIERS IN PHARMACOLOGY
(2021)
Article
Computer Science, Hardware & Architecture
Jens Glaser et al.
Summary: The time-to-solution for structure-based screening of massive chemical databases for COVID-19 drug discovery has been significantly reduced through GPU acceleration and high-throughput optimizations. Over one billion compounds can be successfully docked in a short period of time, providing high-quality compounds for validation experiments.
INTERNATIONAL JOURNAL OF HIGH PERFORMANCE COMPUTING APPLICATIONS
(2021)
Article
Biochemical Research Methods
B. Zagidullin et al.
Summary: This study compares rule-based and data-driven molecular representations in predicting drug combination sensitivity and synergy scores, using standardized results from high-throughput screening studies. The research highlights the importance of supplementing quantitative benchmark results with qualitative considerations for identifying the optimal molecular representation type.
BRIEFINGS IN BIOINFORMATICS
(2021)
Article
Chemistry, Medicinal
Raghad Al-Jarf et al.
Summary: The development of a novel predictive tool, pdCSM-cancer, provides an accurate way to predict molecules likely to be active against one or multiple cancer cell lines by using graph-based chemical structure signatures. This tool includes trained and validated models on data from over 18,000 compounds on 9 tumor types and 74 distinct cancer cell lines, achieving successful predictive performance.
JOURNAL OF CHEMICAL INFORMATION AND MODELING
(2021)
Article
Chemistry, Medicinal
Omar Casanova-Alvarez et al.
Summary: In this study, an automated workflow was developed for predicting antileishmanial activity, utilizing a large, diverse, and highly imbalanced dataset of compounds. The workflow implemented a novel strategy based on balanced training sets and a consensus model using decision trees, resulting in improved predictive accuracy for the test and external sets when compared to other base models like Gaussian-Naive-Bayes and Support-Vector-Machine. The consensus model was found to be effective in prioritizing active compounds with high prediction sensitivity, demonstrating the importance of this approach in QSAR problems.
JOURNAL OF CHEMICAL INFORMATION AND MODELING
(2021)
Review
Economics
Michael Schlander et al.
Summary: The viability of the current commercial R&D model, particularly the cost of developing new drugs, remains a controversial topic. A systematic review of literature found wide variations in estimated R&D costs, with a trend of increasing costs over time, but no clear relationship between cost estimates and study ranking or suitability scores. Future studies should address neglected variables and consider the balance between data transparency and specificity, using proposed suitability scoring systems to assist in addressing issues.
Article
Chemistry, Multidisciplinary
David E. Graff et al.
Summary: Structure-based virtual screening is a crucial tool in early drug discovery for evaluating interactions between target proteins and candidate ligands. Bayesian optimization techniques using surrogate models can significantly reduce computational costs by excluding unpromising compounds from evaluation in large virtual libraries.
Article
Multidisciplinary Sciences
Christoph Gorgulla et al.
Article
Neurosciences
Byung-Hoon Kim et al.
FRONTIERS IN NEUROSCIENCE
(2020)
Article
Chemistry, Medicinal
Douglas E. Pires et al.
JOURNAL OF CHEMICAL INFORMATION AND MODELING
(2020)
Review
Chemistry, Multidisciplinary
Laurianne David et al.
JOURNAL OF CHEMINFORMATICS
(2020)
Review
Pharmacology & Pharmacy
Moe Elbadawi et al.
DRUG DISCOVERY TODAY
(2020)
Article
Chemistry, Medicinal
John J. Irwin et al.
JOURNAL OF CHEMICAL INFORMATION AND MODELING
(2020)
Article
Multidisciplinary Sciences
Hitesh Patel et al.
Article
Chemistry, Multidisciplinary
Timothy E. H. Allen et al.
Review
Chemistry, Multidisciplinary
Tania F. G. G. Cova et al.
FRONTIERS IN CHEMISTRY
(2019)
Article
Mathematical & Computational Biology
Zhihong Liu et al.
DATABASE-THE JOURNAL OF BIOLOGICAL DATABASES AND CURATION
(2019)
Article
Medicine, Research & Experimental
Kiwami Kidana et al.
EMBO MOLECULAR MEDICINE
(2018)
Article
Chemistry, Medicinal
Sabrina Jaeger et al.
JOURNAL OF CHEMICAL INFORMATION AND MODELING
(2018)
Article
Chemistry, Multidisciplinary
Andreas Mayr et al.
Article
Chemistry, Multidisciplinary
Abid Qureshi et al.
JOURNAL OF CHEMINFORMATICS
(2018)
Article
Multidisciplinary Sciences
Othman Soufan et al.
SCIENTIFIC REPORTS
(2018)
Article
Biochemistry & Molecular Biology
Anna Gaulton et al.
NUCLEIC ACIDS RESEARCH
(2017)
Article
Chemistry, Medicinal
Andreas Verras et al.
JOURNAL OF CHEMICAL INFORMATION AND MODELING
(2017)
Article
Chemistry, Medicinal
Stephen J. Capuzzi et al.
JOURNAL OF CHEMICAL INFORMATION AND MODELING
(2017)
Article
Computer Science, Interdisciplinary Applications
Marvin N. Wright et al.
JOURNAL OF STATISTICAL SOFTWARE
(2017)
Article
Chemistry, Multidisciplinary
Egon L. Willighagen et al.
JOURNAL OF CHEMINFORMATICS
(2017)
Article
Chemistry, Multidisciplinary
Jiangming Sun et al.
JOURNAL OF CHEMINFORMATICS
(2017)
Article
Chemistry, Medicinal
Manish Sud
JOURNAL OF CHEMICAL INFORMATION AND MODELING
(2016)
Article
Chemistry, Medicinal
Junshui Ma et al.
JOURNAL OF CHEMICAL INFORMATION AND MODELING
(2015)
Article
Chemistry, Multidisciplinary
V. Ruusmann et al.
JOURNAL OF CHEMINFORMATICS
(2015)
Article
Chemistry, Multidisciplinary
Sereina Riniker et al.
JOURNAL OF CHEMINFORMATICS
(2013)
Review
Biochemistry & Molecular Biology
Jean-Louis Reymond et al.
ACS CHEMICAL NEUROSCIENCE
(2012)
Article
Chemistry, Medicinal
David Rogers et al.
JOURNAL OF CHEMICAL INFORMATION AND MODELING
(2010)
Article
Chemistry, Medicinal
Andreas Bender et al.
JOURNAL OF CHEMICAL INFORMATION AND MODELING
(2009)
Article
Biochemical Research Methods
Justin Klekota et al.
Article
Microbiology
Meredith T. Morris et al.
Article
Biochemistry & Molecular Biology
IV Tetko et al.
JOURNAL OF COMPUTER-AIDED MOLECULAR DESIGN
(2005)
Article
Chemistry, Multidisciplinary
JL Durant et al.
JOURNAL OF CHEMICAL INFORMATION AND COMPUTER SCIENCES
(2002)
Article
Computer Science, Information Systems
MM Breunig et al.