GEM2023

MemCross, a robust computational tool to evaluate membrane permeation coefficients of drug-like molecules

Not scheduled

1h 50m

Poster Clip Session

Speaker

Gabin Fabre (P&T UMR INSERM 1248)

Description

Membrane crossing events are key in the pharmacokinetics of xenobiotics. Among them, passive permeation is the most ubiquitous. Therefore, an accurate and cost-effective prediction of permeation coefficient (logP_erm) is highly valuable. Among the theoretical methods to predict logP_erm, all-atom molecular dynamics simulations are versatile and can provide an accurate description of all intermolecular interactions. However, the cost associated with the required sampling often limits to the study of a few small permeants.
Here, we present MemCross, a tool based on the inhomogeneous solubility-diffusion model and the Accelerated Weight Histogram (AWH) method, and which takes subdiffusion into account. We report the first use of AWH on all-atom membrane simulations. The ease of use of MemCross and its relatively fast convergence allowed to benchmark it on more than 350 xenobiotics (mostly drugs) for which experimental logP_erm are available. We believe this is the largest study performed with drug-like database, while performing all-atom MD simulations, totaling more than 1.2 millisecond of simulation time. A very good correlation was obtained with experimental logP_erm of PC-based liposomes (R² = 0.83). Thus, MemCross is a flexible and affordable tool to evaluate logP_erm of xenobiotics, while providing an atomistic description of the permeation process.