Using Molecular Dynamics to Predict and Understand the Aggregation Behaviour of Small Organic Molecules

Abstract

Finding small-molecule drugs that are effective in the human body has been the main challenge in the drug discovery field for decades. Discovering new candidate inhibitors that exhibit specific binding to a given protein target is challenging in this process. Non-specific inhibition of the protein due to colloidal aggregation of proposed small organic molecules is often problematic, giving false positives in enzyme binding assays.1 To address this problem, rapid data-trained computational filters can be used for identifying and flagging potential aggregate-forming compounds early in drug discovery.2,3 Computational tools have been used to investigate the prediction of aggregators; however, each method has its limitations. This study seeks to address the issue of identifying compound aggregation via the use of molecular dynamics (MD) simulations. A set of 32 diverse small molecules was selected from the Shoichet, Aggregator Advisor and ChemAGG data sets to study their aggregation behaviour. The selection was based on studying a group of compounds that have diverse chemical structures, different net charges, and exhibit a range of aggregation behaviours including non-aggregation. MD simulation was conducted for an initial pool of 23 compounds for 1 µs; a good correlation with the experiment was found in aggregation behaviour. Also, it was observed that the first 100 ns was indicative of the aggregation propensity displayed in the 1 µs trajectories, across the 23 molecules. This finding suggests that 100 ns may be sufficient to predict aggregators from non-aggregators. Application of 100 ns simulations to the remaining 9 compounds found similarly good agreement with the experiment. Various metrics to characterise aggregation during the MD simulations were explored. The metric of average fraction of clusters up to four in percentage (%

Date of Award	31 Dec 2023
Original language	English
Awarding Institution	The University of Manchester
Supervisor	Neil Burton (Supervisor) & Richard Bryce (Supervisor)