Lewis Mervin; Alexey Voronov; Mikhail Kabeshov; Ola Engkvist
ChemRxiv, 2024
https://doi.org/10.26434/chemrxiv-2024-2rlk7-v2

Abstract

Machine-learning (ML) and Deep-Learning (DL) approaches to predict the molecular properties of small molecules are increasingly deployed within the design-make-test-analyse (DMTA) drug design cycle to predict molecular properties of interest. Despite this uptake, there are only a few automated packages to aid their development and deployment that also support uncertainty estimation, model explainability and other key aspects of model usage. This represents a key unmet need within the field and the large number of molecular representations and algorithms (and associated parameters) means it is non-trivial to robustly optimise, evaluate, reproduce, and deploy models. Here we present QSARtuna, a molecule property prediction modelling pipeline, written in Python and utilising the Optuna, Scikit-learn, RDKit and ChemProp packages, which enables the efficient and automated comparison between molecular representations and machine learning models. The platform was developed considering the increasingly important aspect of model uncertainty quantification and explainability by design. We provide details for our framework and provide illustrative examples to demonstrate the capability of the software when applied to simple molecular property, reaction/reactivity prediction and DNA encoded library enrichment analyses. We hope that the release of QSARtuna will further spur innovation in automatic ML modelling and provide a platform for education of best practises in molecular property modelling. The code to the Qptuna framework is made freely available via GitHub.