Explainable AI in drug discovery: self-interpretable graph neural network for molecular property prediction using concept whitening
Open Access
- 31 October 2023
- journal article
- research article
- Published by Springer Nature in Machine Learning
- Vol. 113 (4) , 2013-2044
Abstract
Molecular property prediction is a fundamental task in the field of drug discovery. Several works use graph neural networks to leverage molecular graph representations. Although they have been successfully applied in a variety of applications, their decision process is not transparent. In this work, we adapt concept whitening to graph neural networks. This approach is an explainability method used to build an inherently interpretable model, which allows identifying the concepts and consequently the structural parts of the molecules that are relevant for the output predictions. We test popular models on several benchmark datasets from MoleculeNet. Starting from previous work, we identify the most significant molecular properties to be used as concepts to perform classification. We show that the addition of concept whitening layers brings an improvement in both classification performance and interpretability. Finally, we provide several structural and conceptual explanations for the predictions.Keywords
Funding Information
- Università degli Studi di Roma La Sapienza
This publication has 54 references indexed in Scilit:
- The log P Parameter as a Molecular Descriptor in the Computer-aided Drug Design – an OverviewComputational Methods in Science and Technology, 2012
- A feasible method for optimization with orthogonality constraintsMathematical Programming, 2012
- A Bayesian Approach to in Silico Blood-Brain Barrier Penetration ModelingJournal of Chemical Information and Modeling, 2012
- Quantifying the chemical beauty of drugsNature Chemistry, 2012
- Computational Capabilities of Graph Neural NetworksIEEE Transactions on Neural Networks, 2009
- The Graph Neural Network ModelIEEE Transactions on Neural Networks, 2008
- A priori molecular descriptors in QSAR: a case of HIV-1 protease inhibitorsJournal of Molecular Graphics and Modelling, 2003
- Experimental and computational approaches to estimate solubility and permeability in drug discovery and development settings 1PII of original article: S0169-409X(96)00423-1. The article was originally published in Advanced Drug Delivery Reviews 23 (1997) 3–25. 1Advanced Drug Delivery Reviews, 2001
- Prediction of Physicochemical Parameters by Atomic ContributionsJournal of Chemical Information and Computer Sciences, 1999
- The first general index of molecular complexityJournal of the American Chemical Society, 1981