Machine learning for small molecule drug discovery in academia and industry

Volkamer, Andrea; Riniker, Sereina; Nittinger, Eva; Lanini, Jessica; Grisoni, Francesca; Evertsson, Emma; Rodríguez-Pérez, Raquel; Schneider, Nadine

doi:10.1016/j.ailsci.2022.100056

Cited by 23 publications

(14 citation statements)

References 96 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Large in-house data sets generated with more homogeneous protocols offer the possibility of building high-quality ML models. 46 However, preserving the privacy of molecular structures is of utmost importance in the pharmaceutical industry. To facilitate model building in the public domain, a surrogate data set for CYP inhibition was generated.…”

Section: Assay Variabilitymentioning

confidence: 99%

Deep Learning Models Compared to Experimental Variability for the Prediction of CYP3A4 Time-Dependent Inhibition

Fluetsch,

Trunzer,

Gerebtzoff

et al. 2024

Chem. Res. Toxicol.

Self Cite

View full text Add to dashboard Cite

Most drugs are mainly metabolized by cytochrome P450 (CYP450), which can lead to drug−drug interactions (DDI). Specifically, time-dependent inhibition (TDI) of CYP3A4 isoenzyme has been associated with clinically relevant DDI. To overcome potential DDI issues, high-throughput in vitro assays were established to assess the TDI of CYP3A4 during the discovery and lead optimization phases. However, in silico machine learning models would enable an earlier and larger-scale assessment of TDI potential liabilities. For CYP inhibition, most modeling efforts have focused on highly imbalanced and small data sets. Moreover, assay variability is rarely considered, which is key to understand the model's quality and suitability for decision-making. In this work, machine learning models were built for the prediction of TDI of CYP3A4, evaluated prospectively, and compared to the variability of the experimental assay. Different modeling strategies were investigated to assess their influence on the model's performance. Through multitask learning, additional data sets were leveraged for model building, coming from public databases, in-house CYP-related assays, or other pharmaceutical companies (federated learning). Apart from the numerical prediction of inactivation rates of CYP3A4 TDI, three-class predictions were carried out, giving a negative (inactivation rate k obs < 0.01 min −1 ), weak positive (0.01 ≤ k obs ≤ 0.025 min −1 ), or positive (k obs > 0.025 min −1 ) output. The final multitask graph neural network model achieved misclassification rates of 8 and 7% for positive and negative TDI, respectively. Importantly, the presented deep learning-based predictions had a similar precision to the reproducibility of in vitro experiments and thus offered great opportunities for drug design, early derisk of DDI potential, and selection of experiments. To facilitate CYP inhibition modeling efforts in the public domain, the developed model was used to annotate ∼16 000 publicly available structures, and a surrogate data set is shared as Supporting Information.

show abstract

Section: Assay Variabilitymentioning

confidence: 99%

Deep Learning Models Compared to Experimental Variability for the Prediction of CYP3A4 Time-Dependent Inhibition

Fluetsch,

Trunzer,

Gerebtzoff

et al. 2024

Chem. Res. Toxicol.

Self Cite

View full text Add to dashboard Cite

show abstract

“…In the realm of drug discovery, DL has demonstrated immense potential (Volkamer et al, 2023) due to its ability to process and learn from large and complex data sets (Lavecchia, 2019). Here, we propose a learning pipeline based on Jupyter notebooks for chemists, biologists, and computer scientists alike.…”

Section: Cadd In the Deep Learning Eramentioning

confidence: 99%

TeachOpenCADD goes Deep Learning: Open-source Teaching Platform Exploring Molecular DL Applications

Backenköhler,

Kramer,

Groß

et al. 2023

Preprint

View full text Add to dashboard Cite

TeachOpenCADD is a free online platform that offers solutions to common computer-aided drug design (CADD) tasks using Python programming and open-source data and packages. The material is presented through interactive Jupyter notebooks, accommodating users from various backgrounds and programming levels. Due to the tremendous impact of deep learning (DL) methods in drug design, the TeachOpenCADD platform has been expanded to include an introduction to molecular DL tasks. This edition provides an overview of DL and its application in drug design, highlighting the usage of diverse molecular representations in this field. The platform introduces various neural network architectures, including graph neural networks (GNNs), equivariant graph neural networks (EGNNs), and recurrent neural networks (RNNs). It demonstrates how to use these architectures for developing predictive models for molecular property and activity prediction, exemplified by the Quantum Machine $9$ (QM$9$), ChEMBL, and Kinase Inhibitor BioActivity (KiBA) data sets. The DL edition covers methods for evaluating the performance of neural networks using uncertainty estimation. Furthermore, it introduces an application of GNNs for protein-ligand interaction predictions, incorporating protein structure and ligand information. The TeachOpenCADD platform is continuously updated with new content and is open to contributions, bug reports, and questions from the community through its GitHub repository (\url{github.com/volkamerlab/teachopencadd}). It can be used for self-study, classroom instruction, and research applications, accommodating users from beginners to advanced levels.

show abstract

“…Ideally, deployed tools should be coupled with user interfaces providing capabilities for visualization, data manipulation, and better user experience for chemists and biologists. Moreover, efficient means for feedback loops and timely updates of software tools are needed to further engage users ( Volkamer et al, 2023 ). Platform-as-a-service (PaaS), and Model-as-a-service (MaaS) cloud computing solutions can enhance the monitoring, use, and automated release of in silico solutions.…”

Section: Current Challenges and Future Perspectivesmentioning

confidence: 99%

Cheminformatics and artificial intelligence for accelerating agrochemical discovery

Djoumbou-Feunang,

Wilmot,

Kinney

et al. 2023

Front. Chem.

View full text Add to dashboard Cite

The global cost-benefit analysis of pesticide use during the last 30 years has been characterized by a significant increase during the period from 1990 to 2007 followed by a decline. This observation can be attributed to several factors including, but not limited to, pest resistance, lack of novelty with respect to modes of action or classes of chemistry, and regulatory action. Due to current and projected increases of the global population, it is evident that the demand for food, and consequently, the usage of pesticides to improve yields will increase. Addressing these challenges and needs while promoting new crop protection agents through an increasingly stringent regulatory landscape requires the development and integration of infrastructures for innovative, cost- and time-effective discovery and development of novel and sustainable molecules. Significant advances in artificial intelligence (AI) and cheminformatics over the last two decades have improved the decision-making power of research scientists in the discovery of bioactive molecules. AI- and cheminformatics-driven molecule discovery offers the opportunity of moving experiments from the greenhouse to a virtual environment where thousands to billions of molecules can be investigated at a rapid pace, providing unbiased hypothesis for lead generation, optimization, and effective suggestions for compound synthesis and testing. To date, this is illustrated to a far lesser extent in the publicly available agrochemical research literature compared to drug discovery. In this review, we provide an overview of the crop protection discovery pipeline and how traditional, cheminformatics, and AI technologies can help to address the needs and challenges of agrochemical discovery towards rapidly developing novel and more sustainable products.

show abstract

Machine learning for small molecule drug discovery in academia and industry

Cited by 23 publications

References 96 publications

Deep Learning Models Compared to Experimental Variability for the Prediction of CYP3A4 Time-Dependent Inhibition

Deep Learning Models Compared to Experimental Variability for the Prediction of CYP3A4 Time-Dependent Inhibition

TeachOpenCADD goes Deep Learning: Open-source Teaching Platform Exploring Molecular DL Applications

Cheminformatics and artificial intelligence for accelerating agrochemical discovery

Contact Info

Product

Resources

About