Issue List for 1st Release

[trudack]

This has to be in:

• [x] fix the issue of disappearing C-termini in Rosetta
• [x] mdff and makepsf should have options to add extrabond file and add own restraints
• [x] regenerate chain ID of -result file from mdff run
• [x] regenerate segname for refine output
• [x] cross correlation coloring
• [ ] delete and put in non protein molecules before/after rosetta
• [x] identify densities that have no assigned structures
• [x] optimize clustering for insertion and abinitio analysis
• [ ] adapt clustering model maker selection to VMD atom selections

• [x] regenerate chain ID and segname for full length model

  Would be good to be in:

• analysis option for mdff runs, cross correlation coloring, identify unassigned densities, calculate cc from final structure for everything/an atom selection, calculate cc trajectory, more ?
• include SCWRL4
• include Modeller
• grid scaling based on local resolution
• renaming of chain ID and segname based on txt file with residue range see Issue #14 (double check why some structures only work with -chseg input)
• gap finder with sequence
• step flag for iterative refine
• replace gnuplot by VMD shipped plotting tool (Joao?)

• have VMD to write out mrc files (John)

  Bonus:

• automated refinement strategies based on cross correlation values
• incorporate a module for resolution exchange mdff
• as soon as VMD can write mrc files we can easily automate the cut out of densities to refine only specific areas of the densities or run mdff only on them
• ss structure analysis (Joao)
• helix translate and rotate
• VMD rigid docking to density followed by short MDFF equilibration (John)
• VMD density segmentation of "sausages" by mouse click (John)

[ryanmcgreevy]

I just added the restraints/extrabonds update. Below I have the "has to be in" list which I will cross off when each is finished. Once those are done, we can look into creating individual feature request listings for the others, since those will likely take more time.

• [x] fix the issue of disappearing C-termini in Rosetta
• [x] mdff and makepsf should have options to add extrabond file and also activate/deactive restraints
• [x] regenerate chain ID of -result file from mdff run
• [x] regenerate segname for refine output
• [x] cross correlation coloring
• [x] identify densities that have no assigned structures
• [x] regenerate chain ID and segname for full length model

[ryanmcgreevy]

@trudack it looks like mdff already keeps the chain and segname of the input structure.

I was wrong, this is an issue. As discussed with @trudack, since none of the actual numbering changes, I can regenerate it here just using resids. Will update when this is complete.

Also (note to self), for other naming issues, we can do something similar. refine doesn't change numbering, and full_length_model only works on one chain/segid anyway so you can regenerate based on new segname groups.

[trudack]

@ryanmcgreevy: no only the segnames are right. I assume the -result.pdb is the final output from mdff which is supposed to have the right chain ID and segname. The segnames are OK however the chain IDs are always the first letter of the segname and not the correct ones. Maybe you have an example where by accident this is right or I am looking at the wrong file.

[ryanmcgreevy]

While investigating using VMD clustering instead of the default Rosetta clustering we are using now, I encounter an error with this line in the code (the a variable is undefined when trying to be used later, because there is no match I presume). @maxscheurer can you remember what the purpose of this is? I am pretty unfamiliar with this part of the code...

[maxscheurer]

I think the part/line you're talking about was created to extract the number from the pdb filename listed in the .sc file. We could also just sequentially loop from 1 to nstructs or sth. but sometimes during folding, structures do not converge or so and are subsequently missing in the output file directory. To prevent errors, we thus read the file names from the .sc file to make sure that we only read the files that are really there.

Probably one does not need this annoying regex there and we can just do the loop, probe whether the respective file exists and just skip a number if the file is not there...

[ryanmcgreevy]

@trudack and @maxscheurer , am I correct in thinking that the rosetta_scoring_new.tcl is actually unused and is replaced by rosetta_scoring_wrapper.tcl? I'm trying to clean up everything for an upcoming release and am deleting old files.

[maxscheurer]

Correct. 👍