NMA with EM-data using pseudoatoms

[DavidHaselbach]

Hi,

I am trying to do an NMA with an Electron-microscopy-map. Therefore I produce a pseudo-atom map (as pdb file)
this looks like:

REMARK xmipp_convert_vol2pseudo
REMARK fixedGaussian 8.000000
REMARK intensityColumn Bfactor
ATOM 1 DENS DENS 1 -11.600 -31.600 4.800 1 0.01 DENS
ATOM 2 DENS DENS 2 -43.600 8.400 -27.600 1 0.04 DENS
ATOM 3 DENS DENS 3 -14.400 20.400 16.400 1 0.01 DENS
ATOM 4 DENS DENS 4 -39.200 0.000-120.000 1 0.06 DENS
ATOM 5 DENS DENS 5 -35.600 -3.200 -32.000 1 0.05 DENS
ATOM 6 DENS DENS 6 -24.000 28.400 4.000 1 0.02 DENS
ATOM 7 DENS DENS 7 88.400 -2.800 -76.000 1 0.01 DENS
ATOM 8 DENS DENS 8 -40.000 16.000 -24.000 1 0.07 DENS
ATOM 9 DENS DENS 9 24.000 -43.600 -75.200 1 0.01 DENS
ATOM 10 DENS DENS 10 -4.000 0.000 -75.600 1 0.02 DENS

Unfortunatly I can not parse that file. I get as error, that this is not a valid pdb file or identifier. Is it possiple to somehow work with those files?

Cheers

David

[chakracs]

do the pseudo-atom maps (as PDB files) have a 'standard' format, like the one you show here?

[DavidHaselbach]

yes exactly. That is the standard format

[cihankayacihan]

We will add a parser for EM map standard format.

[jamesmkrieger]

The problem is probably that the residue name should be DEN not DENS because the PDB file format says you need 3 columns for that.

[jamesmkrieger]

Any update on this?

[jamesmkrieger]

We will look into this more at some point I guess.

[jamesmkrieger]

We now have our own tool within ProDy for performing ANM NMA using beads fitted to cryo-EM maps using the topology representing network algorithm. See http://prody.csb.pitt.edu/tutorials/cryoem_tutorial/

Nevertheless, it would be good to handle this output from XMIPP. @ajimoreno, maybe you can help with this.

[jamesmkrieger]

I found the relevant XMIPP code at https://github.com/I2PC/xmipp/blob/devel/src/xmipp/libraries/reconstruction/volume_to_pseudoatoms.cpp and it still has this format so we probably should provide support for it at some point.

[jamesmkrieger]

I think this is probably a niche format, which doesn't really conform to PDB format at all. The fields do not even start at the right columns. For example, field 2 starts at column 6 (starting from 1), which is supposed to be part of the first field so the first field is processed as "ATOM 1". We probably need a completely new parser for it and I don't think that's very far up our priority list right now. I plan to do more ProDy/XMIPP integration next year so probably this will be part of that.

[jamesmkrieger]

I think the way to fix that is on the xmipp side and I think I've found a way to do that

[jamesmkrieger]

There is a parser for this in scipion-em-continuousflex that comes from Situs ModeHunter that could be copied, but I think it's better to change xmipp and continuousflex instead.