MLLPA: A Machine Learning‐assisted Python module to study phase‐specific events in lipid membranes

Abstract

Machine Learning‐assisted Lipid Phase Analysis (MLLPA) is a new Python 3 module developed to analyze phase domains in a lipid membrane based on lipid molecular states. Reading standard simulation coordinate and trajectory files, the software first analyze the phase composition of the lipid membrane by using machine learning tools to label each individual molecules with respect to their state, and then decompose the simulation box using Voronoi tessellations to analyze the local environment of all the molecules of interest. MLLPA is versatile as it can read from multiple format (e.g., GROMACS, LAMMPS) and from either all‐atom (e.g., CHARMM36) or coarse‐grain models (e.g., Martini). It can also analyze multiple geometries of membranes (e.g., bilayers, vesicles). Finally, the software allows for training with more than two phases, allowing for multiple phase coexistence analysis.

LiPyphilic: A Python Toolkit for the Analysis of Lipid Membrane Simulations

“…The modularity of LiPyphilic, along with its focus on integrating with the wider scientific Python stack, means the output of other analysis tools such as FATSLiM or MLLPA can be used as input for further analysis in LiPyphilic. Further, the output of LiPyphilic is in the form of NumPy arrays, Scipy sparse matrices, or Pandas Dataframes.…”

“…If the phase of each lipid at each frame is known, can be used to calculate the registration of L o or L d domains over time. There are various approaches to determining the phase of lipids, from simple metrics such as the deuterium order parameter to more powerful machine learning methods such as hidden Markov models, ,,, Smooth Overlap of Atomic Positions, or those employed by MLLPA . If the lipid phase data are stored in a two-dimensional NumPy array of shape ( N lipids , N frames ), it can be used to create a boolean mask that will tell which lipids to include in the analysis.…”

“…As these inputs are generic NumPy arrays instead of types specific to LiPyphilic, it is possible to use the output from other membrane analysis tools as input to LiPyphilic. For example, you may assign lipids to leaflets using FATSLiM, 3 determine their phase state using MLLPA, 7 or calculate local membrane normals using Mem-Surfer, 4 and extract the results to perform further analysis with LiPyphilic.…”

“…There are various approaches to determining the phase of lipids, from simple metrics such as the deuterium order parameter to more powerful machine learning methods such as hidden Markov models, 29,30,36,56 Smooth Overlap of Atomic Positions, 57 or those employed by MLLPA. 7 If the lipid phase data are stored in a two-dimensional NumPy array of shape (N lipids , N frames ), it can be used to create a boolean mask that will tell Registration which lipids to include in the analysis. For example, if our array is named lipid_phase_data and its elements are strings of either "Lo" or "Ld", then we can select the L o lipids for analysis by passing the boolean mask lipid_pahse_data == "Lo" to the filter_by argument of Registration.…”

“…PyLipID and ProLint are designed for the easy and efficient analysis of lipid–protein interactions. MLLPA is a recently developed Python package that employs various machine learning algorithms to identify the phase L o or L d of lipids in a bilayer . LOOS, on the other hand, is a C++ library with a Python interface for analyzing MD simulations. , Unlike the above packages, LOOS handles the trajectory reading internally while also offering a large set of analysis tools, some of which are for lipid membranes.…”