Enhanced Pathfinding

Important: paths generated with the previous generation will not transfer to the new one. Although you can still open them, the paths will look nothing like before and will need to be deleted and re-calculated. Prolines were also misbehaving slightly; existing RoPE directories will benefit from a run of "Refine all" from the entity menu.

This is a pretty large update.

Features

• Paths have changed significantly (beta feature!):
  • Each calculation of a path is now different. This is randomly seeded by shuffling the order of refinement of the individual bonds. This allows the potential path landscape to be sampled. The path generation can be set to save and auto-restart in order to generate a group.
  • The path calculation is now more robust by running the calculation on improving atom-atom pair distances rather than single-atom trajectories, which also improves the treatment of large motions e.g. full domain swings. The distance threshold can be tuned.
  • The path calculation now uses sine waves instead of a polynomial to naturally match the circular nature of torsion angle changes.
  • After refinement of the backbone, side chain torsion angles are refined to minimise the Lennard-Jones potential between nearby atoms, to better treat Van der Waals contacts. The contact distance threshold can be tuned.
  • Better choice of torsion angle flips has been implemented. Occasionally "protein pirouettes" still occur, and for this reason there is an option to increase the number of combination flip trials that will be carried out to remove them. (default 0)
  • The linearity ratio has been removed as it turned out to be misleading, and has been replaced by the Van der Waals energy activation barrier, which is more meaningful.
  • Rudimentary support for clustering on all pathways of the same type (i.e. sets of paths which all start and end at the same protein instances).
• Tabulated data view within the software, particularly useful for metadata so you can see what you're working with. This also allows you to sort by the values of a particular column and export the reordered table as a CSV.
• The axis explorer has been provided with a few new features, some of which are accessible from a new menu in the top right:
  • the calculation of torsion hotspots has been tweaked, and the table corresponding to the tweaks has been made available. These are calculated as the square root of the sum of the squares of torsion angles spanning the main chain associated with each residue ID.
  • you can export a PDB file at a particular point on the slider.
  • axis limits can be temporarily changed from the default of -1 to +1.
  • hover over an atom to read its description.
• better treatment of hydrogens. Hydrogenated and non-hydrogenated PDB files now produce identical torsion angles. This was previously not the case due to hydrogens being included in the initial alignment process. The RoPE space also now discards torsion angles featuring hydrogen atoms, which means there is no systematic separation between PDB files with hydrogen torsion angle-data and those without.

Bug fixes/minor points

• prolines were misbehaving; existing RoPE directories will benefit from a run of "Refine all" from the entity menu.
• clustering on normalised B factors is now available (beta)
• Ctrl+V (on native apps, Linux and Mac) will allow you to paste into RoPE, making many things easier to transfer.
• Non-ASCII characters may not display, but they will no longer cause a crash.
• depth cueing is now available on protein displays. The slabbing can be altered using keys D/F, following the coot convention.
• all sorts of small improvements to the GUI.