An AI Modeling and Structural Clustering Approach to QAC Disinfectant Development

Abstract

Quaternary ammonium compounds (QACs) have been used as disinfectants for nearly a century due to their broad-spectrum antimicrobial activity and low eukaryotic cytotoxicity. However, their extended use has led bacteria to develop resistance to these compounds through cellular mechanisms such as efflux pumps and cell membrane modification. The proliferation of resistance demonstrates a pressing need for the development of novel disinfectants, as resistant microbes present a threat to global public health. In pursuit of this goal, we have created an artificial intelligence/machine learning model that leverages both antimicrobial data and synthetic insight to generate novel QAC structures, which have been rendered as “clusters” of similar molecular architecture. We prepared 20 exemplars of one such structural cluster and observed broad and potent antimicrobial activity against a suite of Gram-positive, Gram-negative, and resistant bacterial strains. Three of the prepared compounds – Pyr-11(4)12, Pyr-12(4)12, and Pyr-10(4)14 – displayed single-digit micromolar MIC values against this 8-bacteria panel, validating this approach and providing an avenue for future disinfectant design.