Improved graph‐based multitask learning model with sparse sharing for quantitative structure–property relationship prediction of drug molecules

Abstract

The quantitative structure–property relationship (QSPR) is a fundamental technique for evaluating and screening potentially valuable molecules in the field of drug discovery. There is an urgent need to speed up pharmaceutical research and development and a huge chemical space to explore, which necessitate effective and precise computer‐aided QSPR modeling methods. Previous studies with various deep learning models are limited because they are trained on separate small datasets, known as the small‐sample problem. Using transfer learning, this article describes a sparse sharing method that uses advanced graph‐based models to construct an efficient and reasonable multitask learning workflow for QSPR prediction. The proposed workflow is systematically and comprehensively tested with four benchmark datasets containing different targets, and several precisely predicted molecular examples are illustrated. The results demonstrate that an obvious improvement in the prediction of molecular properties is achieved, along with the ability to predict multiple properties simultaneously.

In Silico Prediction of the Temperature-Dependent Decomposition Rate Coefficients of Symmetrical Azo Compounds

“…Alternatively, classical computational approaches, including quantum chemical calculation (QCC) − and kinetic model-based regression, have been successfully employed for calculating different rate coefficients of polymerization reactions (e.g., propagation rate coefficients ( k p ), activation/deactivation rate coefficients, and RAFT equilibrium constants). Attractively, data-driven computational approaches, , such as quantitative structure–property relationship (QSPR) modeling and machine learning (ML) algorithms are becoming an effective tool to predict molecular properties and kinetic parameters. − Recently, a so-called k p (T, NI)-QSPR model was developed by our group to calculate k p values for a wide range of monomers . Junkers et al developed a predictive machine learning model for calculating the k p values of (meth)­acrylates with linear and branched structures, in which both absolute rate coefficients and Arrhenius parameters are predicted with good accuracy .…”