The role and choice of molecular descriptors for predicting retention times in HPLC: A comprehensive review

This review explores the essential role of molecular descriptors (MDs) and their selection for prediction modelling in the domain of high-performance liquid chromatography (HPLC). Currently, there are no standardized methods for selecting MDs, and there is a general lack of understanding about their impact on chromatography, which is more challenging given the multitude of available descriptors. This review aims to provide a comprehensive overview of the role of feature selection methods and an aid for the reader to navigate through MDs in the field of HPLC. It critically assesses the advantages and limitations of the methodologies used since the understanding of more advanced machine learning models. Furthermore, it evaluates the most influential MDs in HPLC and their relationship to retention time, advocating for pursuing innovative descriptor research, using cutting-edge approaches, and interdisciplinary collaboration to surmount challenges and enhance the quality of predictive models in the field of liquid chromatography.