[MDAKit / Hackathon] Secondary Structure Analysis

[IAlibay]

Overview

Secondary structure is a fundamental part of protein conformation.
Algorithms such as STRIDE [Frishman 1995] and
DSSP [Kabsch 1983] calculate backbone dihedral angles
and hydrogen-bonding energy from atomic coordinate to assign secondary
stucture based on a database of assignments.

MDAnalysis does not currently feature a tool for the identification
of secondary structure features (see issues
#1612 and
#2608).
The ability to make atomselections based on the calculated
secondary structure could also be of great use.

Where to start

• Become familiar with the algorithms below and investigate
  external tools that already implement these (e.g.
  mkdspp,
• Add a SecondaryStructure class, using AnalysisBase, that uses
  DSSP and/or STRIDE, implemented natively or using an external
  tool as a dependency, to calculate the secondary structure of a
  selection in each frame.
• Consider allowing the calculated secondary structure to be read
  back alongside the trajectory e.g. using the Auxilliary
  module,
  and allow for atom selections based on secondary structure.

References

1. Frishman,D & Argos,P. (1995) Knowledge-based secondary structure assignment. Proteins: structure, function and genetics, 23, 566-579.
2. Kabsch, W. & Sander,C. (1983) Dictionary of protein secondary structure: pattern recognition of hydrogen-bonded and geometrical features. Biopolymers, 22: 2577-2637.