Development of synthetic chloride transporters using high-throughput screening and machine learning

Abstract

The development of synthetic compounds capable of transporting chloride anions across biological membranes has become an intensive research field in the last two decades. Progress is driven by the desire to develop treatments for chloride transport related diseases (e.g., cystic fibrosis), cancer or bacterial infections. In this manuscript, we use high-throughput screening and machine learning to identify novel scaffolds, and to find the molecular features needed to achieve potent chloride transport that can be generalized across diverse chemotypes. 1894 compounds were tested, 59 of which had confirmed transmembrane chloride transport ability. A machine learning (ML) binary classification model indicated that MolLog P is the most important feature to predict transport ability, but it is not sufficient by itself. The best ML model was able to identify potential chloride transporters from the DrugBank database and the predictions were experimentally validated. These insights can provide other researchers with inspiration and guidelines to develop ever more potent chloride transporters.