Reorganising the CCP4i modules and tasks

CCP4i MR module reorganisation meeting

5th April 2006

Peter Briggs and Ronan Keegan discussed possible improvements to the CCP4i MR modules organisation based on comments from Charlie Bond at the CCP4 Developers Meeting.

The aim of the discussion was to determine what improvements could be made to the existing layout and using the existing tasks as much as possible. It seems that some kind of two-tier rearrangement (also importing some relevant tasks from other modules) could greatly enhance the usability of the interface in this area. A suggested "dream MR module" would consist of the following components and arrangement:

Top level Subtasks
Analysis Cell content analysis
Analyse data for MR
Self RF in polars
Model Generation Chainsaw ("Create search model")
Modeller ("Edit protein structure")
Molrep model generation
MRBump model generation (uses Chainsaw, Molrep, PdbSet/Cur (polyalanine/pdbclip)
Phaser NMA?
Beast (deprecated)
Amore Amore Model Database
Generate SFs for MR
Edit Amore Solution
Automated MR MRBump
Utilities Edit PDB File (some functions useful for making models)
Superpose (for Phaser model preparation)
Import/edit Protein Sequence

In this table red represents an existing task, blue represents a task that doesn't exist at present and green represents a task that could be made from an existing task with some modifications.

Some other comments arose from the discussion:

  1. Although the Amore MR database is probably not viable for extension to other MR tasks, RMK felt that some sort of "MR model manager" might be useful.
  2. The status of the Phaser NMA (normal modes analysis) function is unclear, in particular how does it generate the model?
  3. For the green and blue highlighted tasks, it might be possible to reuse existing tasks in the short term by having some mechanism to launch the tasks with modified defaults: for example, MrBUMP could be started with the "model generation only" option pre-selected when launched from the "Model generation" submenu. This would require some changes to the CCP4i core.
  4. It would be useful in future to combine all the diverse analysis tasks into a single MR analysis task. Along similar lines Charlie Bond has suggested that a general analysis task could be run e.g. as part of the Data Reduction module.
  5. It is not clear what the best way to represent the two-tier arrangement will be in practice within CCP4i. A number of possible options were suggested, including: "folder"-like arrangement (cf CCP4i task windows); cascading menus (cf Windows programs menu); "browser"-like model (clicking on e.g. the Analysis button replaces the view with the submenu). It is clear that some market research will be needed.
  6. It was felt that the "core" MR programs (Phaser, Molrep, Amore and Beast) would be best presented as being at the top-level (as shown above) rather than in a submenu. However this is also open to discussion.
  7. There is also the question of whether MrBUMP should be one of the core MR tasks, or whether it's sufficient to leave have it only in the "Automated Tasks" section.

RMK will be visiting Charlie Bond on 26/27th April so it would be good to have at least a prototype to show. PJB will look at hacking together a demo version for RMK to take; RMK is to remind PJB to do this before he goes.

Suggestions for reorganising the Experimental Phasing module from Charlie Bond

4th May 2006

Email from Charlie Bond following Ronan's visit to Dundee suggested a similar rearrangement of the tasks in the Experimental Phasing module of CCP4i.

Top level Subtasks
Data Preparation Combine and Scale datasets 1
Prepare and Normalise (Revise)
Automated Phasing SHELX C/D/E
Heavy Atom Location Tools Patterson 2
Phasing Acorn
Visualisation Coot/MG view maps + sites
MapSlicer view maps + sites
Utilities CAD
FHSCAL (actually part of Scaleit task?)

Charlie's Notes

1. Merge and Scale are currently separate tasks - this seems pointless, if you merge them , you want to scale them. Put differently why not allow multiple files/columns as input to the scaling task.

2. Patterson could go under "Utilities" if a Patterson-calculating step is prepended to the RSPS task - this is probably a sensible thing to do anyway.

Suggestions for reorganising the Refinement module from Charlie Bond

8th June 2006

Email from Charlie Bond suggesting a rearrangement of the tasks in the Refinement module of CCP4i.

Top level Subtasks
Model Preparation Coot
Convert Coordinate Formats
Edit PDB file
[Download PDB file and SFs from PDB]
Restraint Preparation Edit Restraints
Merge Monomer Libraries
Create/Edit TLS
[Define NCS]
Refinement Refmac
NCS Phased Refinement
Model Completion and Analysis Add/Tidy Waters
Analyse TLS *** 1
Analyse ANISOU *** 1

Charlie's Notes

1. "***" indicates that these tasks should perhaps be run automatically after refinement. Why not do PROCHECK/Molprobity every time? If you used TLS or had ANISOUs, then why not do the analysis automatically.

Peter Briggs 13/6/2006