Molecular dynamics simulations as a guide for modulating small molecule aggregation

Area of Science:

• Computational Chemistry
• Drug Discovery
• Biochemistry

Background:

• Small colloidally aggregating molecules (SCAMs) pose challenges in biological assays during drug discovery.
• Understanding and predicting SCAMs' self-associating properties is crucial for drug delivery and analytical biochemistry.
• Current chemoinformatics filters (ChemAGG, Aggregator Advisor) have limitations due to training data quality and diversity.

Purpose of the Study:

• To evaluate molecular dynamics (MD) simulations as a physics-based method for predicting small organic molecule aggregation propensity.
• To compare the accuracy of MD simulations against existing chemoinformatics filters for aggregation prediction.
• To explore structure-aggregation relationships and identify chemical modifications to modulate aggregation behavior.

Main Methods:

• Performed 100 ns MD simulations in explicit solvent for a set of 32 diverse molecules.
• Utilized implicit solvent simulations with the generalized Born model for comparative analysis.
• Analyzed simulation data to assess aggregation propensity and identify structure-activity relationships.

Main Results:

• MD simulations achieved a 97% success rate in predicting aggregation propensity, significantly higher than Aggregator Advisor (75%) and ChemAGG (72%).
• Short-timescale MD simulations effectively captured dynamic aggregation behaviors, comparable to longer microsecond trajectories.
• Implicit solvent models were less effective in predicting aggregation compared to explicit solvent MD simulations.

Conclusions:

• MD simulations offer a robust, physics-based approach for predicting small molecule aggregation propensity across diverse chemical structures.
• MD simulations provide valuable insights into structure-aggregation relationships, guiding compound optimization.
• While lower throughput than chemoinformatics filters, MD simulations are suitable for focused subsets and offer detailed guidance for modifying aggregation behavior.