segmentation.geostas

Bio3D is an R package that encapsulates different tools for biological structure analysis. The GeoStaS domain finder tries to locate rigid domains based on trajectory movements. The generated files come from a custom script written in R, so their format is controlled by our workflow rather than by the external project itself.

Documentation can be found at: http://thegrantlab.org/bio3d/reference/geostas.html

Module Contents

class segmentation.geostas.Parser(pdb_path, clustering_directory_path)

Bases: segmentation.SegmentationParser

The base class for all segmentation parsers.

parse() → collections.abc.Iterator[Tuple[str, int, str]]

The return value is a collection of items of the form:

("<method name>", <number of domains>, "<chopping>")

They could be collected into a list or yield-ed. The write_segmentations() function only expects that this method returns an iterable object.

_translate_atoms_to_residues(atom_data, atom_groups)

The output of GeoStaS is (alpha carbon) atom indices (1-indexed) while the output of the parser is in residues. This method translates the sequential atom indices into the residues they correspond to by using an MDAnalysis universe.

_generate_chopping(residue_groups)

Input: [[1, 2, 3], [10, 11, 20, 21], ...] Output: 1-3,10-11_20,21,...