Article
Multidisciplinary Sciences
Hiromasa Kaneko
Summary: This study evaluated the prediction accuracy of models constructed using selected X and investigated the results of variable selection methods. The findings suggest that even when variables unrelated to y are selected, accurate models can be constructed by applying various regression analysis methods.
Review
Engineering, Biomedical
Jacob Kerner, Alan Dogan, Horst von Recum
Summary: Machine learning has been widely utilized in various fields, including biomaterials, optimizing data collection and analysis. Recent advances in biomaterials have focused on quantitative structure properties relationships, introducing four basic models for rapid development and addressing the lack of machine learning implementation in the field. This article aims to spark greater interest and awareness in utilizing computational methods for biomaterials research.
ACTA BIOMATERIALIA
(2021)
Review
Polymer Science
Yuankai Zhao, Roger J. Mulder, Shadi Houshyar, Tu C. Le
Summary: Polymers are important materials with diverse properties, and machine learning has shown great potential in data-driven polymer design. Machine learning models trained on polymer datasets can accurately predict polymer properties and assist in candidate polymer screening before lab synthesis.
Article
Environmental Sciences
Juan Jose Villaverde, Beatriz Sevilla-Moran, Jose Luis Alonso-Prados, Pilar Sandin-Espana
Summary: Any active substance with phytosanitary capacity intended to be marketed in Europe must pass exhaustive controls to assess its risk before being marketed and used in European agriculture. In this research work, the open literature collection on alloxydim was used to propose potential chlorination paths from alloxydim isomers. Furthermore, several QSAR/QSPR models have been used to fill the of knowledge gap relative to some key parameters in the physico-chemical, environmental and ecotoxicological areas of potential alloxydim TPs from chlorinated water for which little information exists. These and other results highlight that the hazards of several TPs should be seriously considered and reopen the debate on the implications of the use of QSAR/QSPR models for pesticide risk assessment in the legislative framework.
SCIENCE OF THE TOTAL ENVIRONMENT
(2022)
Review
Biochemistry & Molecular Biology
Yasunari Matsuzaka, Yoshihiro Uesawa
Summary: The QSAR approach has been widely used for a long time, with the development of deep learning methods showing excellent performance in modeling, especially in the field of classification. Proper feature extraction and selection are crucial for building models with outstanding performance.
CURRENT ISSUES IN MOLECULAR BIOLOGY
(2021)
Article
Chemistry, Multidisciplinary
Hiromasa Kaneko
Summary: This paper proposes a method based on SELFIES for molecular descriptors, structure generation, and inverse QSAR/QSPR. By converting SELFIES into SELFIES descriptors x, an inverse analysis of the QSAR/QSPR model y = f(x) is conducted, obtaining x values that achieve the target y value and successfully generating SELFIES strings or molecules.
Article
Chemistry, Medicinal
Alexander Y. Sedykh, Ruchir R. Shah, Nicole C. Kleinstreuer, Scott S. Auerbach, Vijay K. Gombar
Summary: Molecular structure-based predictive models offer a cost-effective and efficient alternative to animal testing, with the need for interpretable descriptors to provide chemistry-backed predictive reasoning. Saagar, a novel chemistry-aware substructure, outperformed publicly available fingerprint sets in extracting compounds with higher scaffold similarity, showcasing its ability to efficiently characterize diverse chemical collections.
CHEMICAL RESEARCH IN TOXICOLOGY
(2021)
Article
Computer Science, Artificial Intelligence
Fan Zhang, Jianshen Zhu, Rachaya Chiewvanichakorn, Aleksandar Shurbevski, Hiroshi Nagamochi, Tatsuya Akutsu
Summary: The paper introduces a method for computer-aided drug design using intelligent systems, with a focus on the inverse QSAR/QSPR approach. It presents the two phases of the method, forward prediction and inverse inference, and proposes a new method for inferring acyclic chemical compounds. Computational experiments show that the proposed method outperforms existing methods.
APPLIED INTELLIGENCE
(2022)
Article
Chemistry, Physical
Mina Najafi, Ali Reza Akbarzadeh, Rahmatollah Rahimi, Mohammad Hossein Keshavarz
Summary: A novel QSPR model was presented to predict the photodegradation average rate of 4-NP and 4-CP by biomolecules. The model achieved high accuracy by calculating metal interaction with ligand and fragment effects, using measured values of 70 porphyrin complexes.
JOURNAL OF MOLECULAR STRUCTURE
(2022)
Article
Chemistry, Medicinal
Joseph Bloxham, Daniel Hill, Neil F. Giles, Thomas Allan Knotts, W. Vincent Wilding
Summary: Quantitative Structure-Property Relationships (QSPRs) are widely used in chemistry and engineering for predicting complex molecular properties. However, existing QSPRs for liquid heat capacity (cpl) are limited to single temperatures and have issues with oxygen-containing functional groups.
MOLECULAR INFORMATICS
(2022)
Article
Chemistry, Medicinal
Werner J. Geldenhuys, Jeffrey R. Bloomquist
Summary: The research aimed to develop cheminformatic models to describe brain uptake in a novel insect model and evaluate the predictive ability.
BIOORGANIC & MEDICINAL CHEMISTRY LETTERS
(2021)
Article
Environmental Sciences
Flavio O. Sanches-Neto, Jefferson R. Dias-Silva, Vitor M. de Oliveira, Vincenzo Aquilanti, Valter H. Carvalho-Silva
Summary: The rate constants of OH reactions with atmospheric organic pollutants are important parameters for understanding the kinetics and mechanisms of these reactions. This study developed a protocol based on machine learning and molecular fingerprints to estimate the rate constants, and the results showed that the model had a high predictive capacity. Additionally, the contribution of chemical classes to reaction kinetics and mechanism was analyzed using the SHAP method.
ATMOSPHERIC ENVIRONMENT
(2022)
Article
Pharmacology & Pharmacy
Mare Oja, Sulev Sild, Geven Piir, Uko Maran
Summary: In this study, authors developed data-driven models for predicting intrinsic aqueous solubility of drug substances using curated training data sets and derived three quantitative structure-property relationships. The models, which are mechanistically transparent and easy to understand, showed significant improvement in prediction capability and reduction of outliers through a consensus modeling approach. The developed models have been published in the QsarDB.org repository according to FAIR principles for unrestricted use in exploration, downloading, and predictions.
Article
Mathematics, Applied
Juan Rada, Jose M. Rodriguez, Jose M. Sigarreta
Summary: This paper introduces a topological index called the Sombor index, which is a molecular descriptor that has received considerable research attention in recent years. The paper also proposes a family of topological indices called integral Sombor indices, which generalize the Sombor index, and explores their application in QSPR/QSAR research.
APPLIED MATHEMATICS AND COMPUTATION
(2023)
Article
Automation & Control Systems
Hiromasa Kaneko
Summary: Support vector regression (SVR) can capture the nonlinear relationship between explanatory variables X and a target variable y, leading to high predictive accuracy. The novel VI-SVR method, which considers variable importance, outperforms traditional SVR in predictive accuracy.
JOURNAL OF CHEMOMETRICS
(2021)
Article
Chemistry, Medicinal
Ulf Norinder, Ola Spjuth, Fredrik Svensson
JOURNAL OF CHEMICAL INFORMATION AND MODELING
(2020)
Article
Chemistry, Multidisciplinary
Francesco Gentile, Vibudh Agrawal, Michael Hsing, Anh-Tien Ton, Fuqiang Ban, Ulf Norinder, Martin E. Gleave, Artem Cherkasov
ACS CENTRAL SCIENCE
(2020)
Article
Chemistry, Medicinal
Marina Garcia de Lomana, Andrea Morger, Ulf Norinder, Roland Buesen, Robert Landsiedel, Andrea Volkamer, Johannes Kirchmair, Miriam Mathea
Summary: Computational methods such as machine learning approaches have shown success in predicting in vitro outcomes, but their ability to predict in vivo endpoints is more limited. Recent studies suggest that combining chemical and biological data can lead to better models for in vivo endpoints.
JOURNAL OF CHEMICAL INFORMATION AND MODELING
(2021)
Article
Chemistry, Medicinal
Jin Zhang, Ulf Norinder, Fredrik Svensson
Summary: Predictive modeling for toxicity, especially when combining deep learning with the conformal prediction framework, can lead to highly predictive models with well-defined uncertainties. This approach shows promising results on Tox21 challenge data, delivering toxicity predictions with confidence and statistically better performance on minority class predictions compared to underlying models.
JOURNAL OF CHEMICAL INFORMATION AND MODELING
(2021)
Article
Chemistry, Organic
James J. Douglas, Matthew R. Tatton, Daniel de Bruin, David Buttar, Calum Cook, Kuangchu Dai, Catalina Ferrer, Kevin Leslie, James Morrison, Rachel Munday, Thomas O. Ronson, Hucheng Zhao
Summary: In this study, route design and synthesis of a challenging chirally atropisomeric inhibitor were conducted to improve synthesis efficiency and avoid racemization. The strategy was further validated on other substrates.
JOURNAL OF ORGANIC CHEMISTRY
(2022)
Article
Chemistry, Multidisciplinary
Ulf Norinder, Ola Spjuth, Fredrik Svensson
Summary: This study investigates the performance of synergy conformal prediction on bioactivity data and demonstrates its effectiveness in federated learning. The results show that synergy conformal predictors based on randomly sampled training data are competitive, while using completely separate training sets often leads to poorer performance.
JOURNAL OF CHEMINFORMATICS
(2021)
Article
Business
Ulf Norinder, Petra Norinder
Summary: In this study, the combination of deep learning and conformal prediction was used to predict the sentiment of Amazon product reviews. The results showed high accuracy and efficiency in predicting the sentiment of the test set, both within the same category and across different categories. Additionally, the combination of deep learning and conformal prediction was found to handle class imbalances without explicit class balancing measures.
JOURNAL OF MANAGEMENT ANALYTICS
(2022)
Article
Engineering, Environmental
Elena Dracheva, Ulf Norinder, Patrik Ryden, Josefin Engelhardt, Jana M. Weiss, Patrik L. Andersson
Summary: Data on toxic effects of industrial chemicals are lacking in the current understanding. This study developed in silico models using high-throughput screening data to identify potential thyroid hormone system-disrupting chemicals. The models were applied to two different databases, identifying chemicals of concern for thyroid hormone disruption.
ENVIRONMENTAL SCIENCE & TECHNOLOGY
(2022)
Article
Chemistry, Medicinal
Alzbeta Tuerkova, Brandon J. Bongers, Ulf Norinder, Orsolya Ungvari, Virag Szekely, Andrey Tarnovskiy, Gergely Szakacs, Csilla Ozvegy-Laczka, Gerard J. P. van Westen, Barbara Zdrazil
Summary: The integration of statistical learning methods with structure-based modeling approaches is an effective strategy to identify novel lead compounds in drug discovery. In this study, a consensus virtual screening approach combined with molecular docking was used to discover highly active novel inhibitors for hepatic OATPs. The structural differences in ligand binding to the three transporters were explained through structural comparison of the detected binding sites and docking poses.
JOURNAL OF CHEMICAL INFORMATION AND MODELING
(2022)
Article
Chemistry, Medicinal
Frederick W. Goldberg, Jason G. Kettle, Gillian M. Lamont, David Buttar, Attilla K. T. Ting, Thomas M. McGuire, Calum R. Cook, David Beattie, Pablo Morentin Gutierrez, Stefan L. Kavanagh, Jasper C. Komen, Aarti Kawatkar, Roger Clark, Lorna Hopcroft, Gareth Hughes, Susan E. Critchlow
Summary: Due to increased reliance on glycolysis, monocarboxylate transporters (MCTs) are upregulated in cancer. MCT4 inhibition can lead to cytotoxic levels of intracellular lactate and may be of interest for immuno-oncology. A triazolopyrimidine hit was identified as a potential MCT4 inhibitor, and further modifications were made to improve potency, selectivity, and other properties. The resulting clinical candidate 15 (AZD0095) has excellent potency, MCT1 selectivity, clean mechanism of action, suitable properties for oral administration, and good preclinical efficacy.
JOURNAL OF MEDICINAL CHEMISTRY
(2023)
Article
Chemistry, Medicinal
Nadim Akhtar, Marianne B. Ashford, Louisa Beer, Alex Bowes, Tony Bristow, Anders Broo, David Buttar, Steve Coombes, Rebecca Cross, Emma Eriksson, Jean-Baptiste Guilbaud, Stephen W. Holman, Leslie P. Hughes, Mark Jackman, M. Jayne Lawrence, Jessica Lee, Weimin Li, Rebecca Linke, Najet Mahmoudi, Marc McCormick, Bryce MacMillan, Ben Newling, Maryann Ngeny, Claire Patterson, Andy Poulton, Andrew Ray, Natalie Sanderson, Silvia Sonzini, Yayan Tang, Kevin E. Treacher, Dave Whittaker, Stephen Wren
Summary: The recent emergence of drug-dendrimer conjugates presents analytical and measurement challenges in the pharmaceutical industry. These complex molecules have high molecular weights and diverse characteristics. The understanding and definition of their characteristics and quality attributes, which impact efficacy and safety, require measurement of molecular weight, impurity characterization, quantification of conjugated versus free API molecules, determination of impurity profiles, primary structure, particle size, and morphology. This study provides a global characterization of a drug-dendrimer conjugate and discusses the impact of various analytical and measurement techniques on understanding this complex molecular entity. The results are crucial for the future development of dendrimer-based medicines.
JOURNAL OF PHARMACEUTICAL SCIENCES
(2023)
Article
Chemistry, Medicinal
John G. G. Cumming, Lukas Kreis, Holger Kuehne, Roger Wermuth, Maarten Vercruysse, Christian Kramer, Markus G. G. Rudolph, Zhiheng Xu
Summary: Novel bacterial topoisomerase inhibitors (NBTIs) have been discovered to target clinically validated bacterial type II topoisomerases and effectively combat multidrug-resistant Gram-negative bacteria. The discovery of a series of NBTIs with a novel indane DNA binding moiety, as well as their interaction with Staphylococcus aureus DNA gyrase-DNA, has been reported. The lead compound 18c shows potent broad-spectrum activity against multidrug-resistant Gram-negative bacteria.
ACS MEDICINAL CHEMISTRY LETTERS
(2023)
Article
Polymer Science
R. M. England, S. Sonzini, D. Buttar, K. E. Treacher, M. B. Ashford
Summary: This study characterized poly(l-lysine) dendrimers using advanced analytical techniques and molecular dynamics simulations. The results showed an increase in refractive index and intrinsic viscosity in the early generations of the dendrimers, which decreased in later generations. The protected dendrimers had different molecular density profiles compared to the unprotected dendrimers, possibly due to electrostatic repulsion. This research provides valuable insights into the structure and properties of PLL dendrimers for drug delivery applications.
Article
Chemistry, Multidisciplinary
Kjell Jorner, Tore Brinck, Per-Ola Norrby, David Buttar
Summary: The study introduces hybrid models combining traditional transition state modeling and machine learning to accurately predict reaction barriers, offering competitive accuracy in low-data scenarios.
Article
Chemistry, Multidisciplinary
Amol Thakkar, Simon Johansson, Kjell Jorner, David Buttar, Jean-Louis Reymond, Ola Engkvist
Summary: The article discusses the development of synthesis planning technologies and the relevance of computer-assisted synthesis planning (CASP) in drug discovery and development. It emphasizes the need for an automated synthesis platform to enhance chemical workflows, and highlights the interaction between experimental and computational scientists as a key driver of technological development.
REACTION CHEMISTRY & ENGINEERING
(2021)