Predicting Drug Solubility Using Different Machine Learning Methods -- Linear Regression Model with Extracted Chemical Features vs Graph Convolutional Neural Network

TL;DR

This study compares linear regression and graph convolutional neural network models for predicting drug solubility, highlighting the trade-offs between interpretability and performance using chemical features and graph-based representations.

Contribution

It demonstrates the effectiveness of GCNNs over linear regression in prediction accuracy and emphasizes the interpretability advantages of linear models in chemical analysis.

Findings

GCNN outperforms linear regression in prediction accuracy

Oxygen atoms increase molecule solubility

Most heteroatoms decrease solubility except oxygen and nitrogen

Abstract

Predicting the solubility of given molecules remains crucial in the pharmaceutical industry. In this study, we revisited this extensively studied topic, leveraging the capabilities of contemporary computing resources. We employed two machine learning models: a linear regression model and a graph convolutional neural network (GCNN) model, using various experimental datasets. Both methods yielded reasonable predictions, with the GCNN model exhibiting the highest level of performance. However, the present GCNN model has limited interpretability while the linear regression model allows scientists for a greater in-depth analysis of the underlying factors through feature importance analysis, although more human inputs and evaluations on the overall dataset is required. From the perspective of chemistry, using the linear regression model, we elucidated the impact of individual atom species and functional groups on overall solubility, highlighting the significance of comprehending how chemical structure influences chemical properties in the drug development process. It is learned that introducing oxygen atoms can increase the solubility of organic molecules, while almost all other hetero atoms except oxygen and nitrogen tend to decrease solubility.