Characterization of domain formation in complex membranes.

Area of Science:

• Computational Biophysics
• Membrane Biophysics
• Molecular Dynamics Simulations

Background:

• Lipid membranes naturally separate into disordered (Ld) and ordered (Lo) domains.
• This domain separation influences membrane physical and biological processes, including electroporation and protein sorting.
• Advancements in computational power and simulation software enable detailed study of these domains using classical molecular dynamics.

Purpose of the Study:

• To present a versatile and robust analysis pipeline for identifying lipid membrane domains.
• To provide practical guidance and coding examples for implementing the pipeline in molecular dynamics studies.
• To adapt the pipeline for diverse lipid compositions and analyze phase separation tendencies.

Main Methods:

• Utilizes Gaussian-based Hidden Markov Models to predict lipid order states based on area per lipid and S_CC order parameters.
• Employs the Getis-Ord local spatial autocorrelation statistic on a Voronoi tessellation to identify regions of correlated ordered lipids.
• Applies the pipeline to coarse-grained molecular dynamics simulations of specific lipid mixtures (DPPC/DIPC/cholesterol and POPC/PUPC/cholesterol).

Main Results:

• Successfully identifies and quantifies lipid domain separation in simulated membrane systems.
• Demonstrates a strong tendency towards phase separation in a DPPC/DIPC/cholesterol mixture.
• Shows a weak tendency towards phase separation in a POPC/PUPC/cholesterol mixture, highlighting the pipeline's sensitivity to composition.

Conclusions:

• The developed analysis pipeline offers a comprehensive and adaptable workflow for studying lipid membrane domain formation.
• The pipeline aids in understanding the physical and biological implications of lipid domain organization.
• Provides practical, code-driven insights for researchers investigating membrane heterogeneity.