LOOS 4.1.0
The Lightweight Object Oriented Structural analysis library/toolkit
|
A collection of tools for 2D Voronoi decomposition of slabs, presumably from a membrane simulation. This package is a wrapper around the scipy (www.scipy.org) Voronoi class, which is itself a wrapper around QHull (www.qhull.org). As a result, this package is dependent having a working install of numpy and scipy (instructions for installing them can be found at www.scipy.org).
All tools assume work with a slab-like geometry oriented along the z-axis. Within a slab, the system is treated as 2-dimensional.
All tools take at least 2 selections of atoms. The first is the set of atoms used to compute the Voronoi decomposition, while any other selections are subsets of those atoms for which areas are desired. In all cases, these selections are application after the original selection.
Padding: all of the tools listed have a required option called "padding". The purpose of this option is to ensure that periodic boundaries are handled correctly. Since the underlying engine (qhull) has no knowledge of periodicity, we do this by replicating dummy atoms in the 8 surrounding periodic images, such that no "true" atoms are at the "edge". The padding option controls how far out these padding atoms are generated. If your selection for the Voronoi decomposition is all-atom or all heavy-atom, 15 angstroms is a good choice for the padding value. However, if you're using a sparser selection (e.g. just lipid phosphates), you will probably need to make the padding value significantly larger, e.g. 30 ang or more. If your padding value is too small, you will occasionally see absurdly large areas for individual atom, or the program can fail outright.
Compute areas for different sets of atoms within a particular slice along the membrane normal. Typical slice widths are 2-4 ang.
Compute the voronoi cross-sectional area for something (e.g. a protein) through the membrane.