MoDeSuS: A Machine Learning Tool for Selection of Molecular Descriptors in QSAR Studies Applied to Molecular Informatics

Abstract

The selection of the most relevant molecular descriptors to describe a target variable in the context of QSAR (Quantitative Structure-Activity Relationship) modelling is a challenging combinatorial optimization problem. In this paper, a novel software tool for addressing this task in the context of regression and classification modelling is presented. The methodology that implements the tool is organized into two phases. The first phase uses a multiobjective evolutionary technique to perform the selection of subsets of descriptors. The second phase performs an external validation of the chosen descriptors subsets in order to improve reliability. The tool functionalities have been illustrated through a case study for the estimation of the ready biodegradation property as an example of classification QSAR modelling. The results obtained show the usefulness and potential of this novel software tool that aims to reduce the time and costs of development in the drug discovery process.

Leveraging genetic algorithms to maximise the predictive capabilities of the SOAP descriptor

“…As shown in previous work, 22–25 the choice of descriptor has a large impact on the quality of predictions. There is a huge selection of descriptors available to use, and if some thought is not employed to understand which ones are appropriate, it may lead to poor results.…”

Machine learning in materials modeling—fundamentals and the opportunities in 2D materials

“…to predict BOD of chemical compounds [37]. Here we discuss some widely used ML algorithms and their modifications made suitable for materials.…”

mibPOPdb: An online database for microbial biodegradation of persistent organic pollutants

“…However, QSAR models are inevitably associated with drawbacks that limit their application. The reliability of QSAR models and precision of ready biodegradability results is dependent on the correct feature selection applied during QSAR modeling [ 42 ]. Also, when QSAR models are applied to chemicals outside the applicability domain for which they were developed, it results in added conservatism being incorporated and an increase in error propagation within the biodegradability classification model [ 43 ].…”