Development plan for MrBUMP
Go to Ronan's page.
Go to Martyn's page.
Go to the mrbump project management page..
Updated RMK: 3rd May 2007.
- Allow Model Generation only mode to run without an MTZ file.
- Finish port to Windows, i.e. provide an installer (likely to be
a variant of setup.py) (RMK)
- Extension to complexes of more than one kind of protein chain (RMK,
- Use of PISA instead of PQS to generate multimer models (MDW, in progress).
- Include more sequences in the multiple alignment (not necessarily
just those with associated structures) to improve stability / accuracy.
- Adjust Chainsaw so removes one residue either side of an insertion.
- Run Pirate after MR, when can't run Acorn.
- Look at Rapper for generating ensembles and for rebuilding
- Automatically determine members of ensemble. Look at RMSD values
and stop adding pdbs when a threshold is reached. (MDW)
- Allow users to input RMS for specified models.
- Use HHPred for doing sequence-based search.
- Add an automatic update mechanism. This to be in place for CCP4 6.1 release.
- Investigate use of buccaneer and ccp4_pipeline_simple for model building.
Short term plans
- Set up auto downloads when mrbump goes into the 6.1 release.
- Run data validation tools at the beginning, e.g. sfcheck. Check data quality, etc..
- Success criteria different when tNCS present.
- Allow for more than 1 ensemble.
- Removal of uncertain regions (e.g. flexible loops) from
models based on B factors (MDW)
- MrBUMP Preferences stored in ~/.ccp4_mrbump and accessed
by new interface. Include general options (e.g. whether to FASTA search
locally, which multiple alignment program to use, etc.). Specify where
to store database files. Look at this for release 6.1. (RMK)
- Run Buccaneer after MR. Kevin says this should work now.
- Run Arp/warp after MR, for model re-building. (RMK)
- Possible inclustion of Mustang_MR program for generation of additional
ensemble models for MR. (RMK)
Long term plans
- Produce a webservice version.
- Do MR for all possible spacegroups, or a choice of spacegroups.
- Use of local PDB coordinate or Balbes databases.
- Search for more than one domain per target molecule (whether
multiple copies of same domain or different domains)
- Removal of flexible loops based on geometry and/or GNM
calculation and/or hydrophobicity
- Generate models from normal modes analysis
- Better homologue search based on profiles from large sequence
families, e.g. psiBLAST, ScanPS
- Include externally-provided multiple alignment.
- Alternative models based on alternative alignments.
- Allow all models in an NMR structure to be used in MR.