I have a pdb file for an entire virus (biological assembly) that I can read into Chimera just fine, as either a multimodel PDB with 60 different models, or as a PDB lacking model and endmdl cards, but with the atom number of the multi-model pdb preserved.  The model is placed very close to the density for the virus in a tomogram.  When I use the fit in map command, the single model pdb moves nicely by a couple of nm and a few degrees into the tomogram density.  For the multimodel map I must first use the combine command in the model panel to combine the 60 differernt models, but that then also moves nicely into the density.  However, in both cases, when I try to write out the slightly rotated/translated coordinates to a PDB file, I wind up with a semi-corrupted PDB file containing **** in many locations, especially for the last half of the file. Since the virus is pseudo T=31 with 18 chains in the viral asu, the pdb model has more than 2 million atoms and 1048 chains; the chains obviously cannot be uniquely identified with a single letter, the atom numbers exceed the the allowed format, and residues are renumbered and then also exceed the allowed format.

I can fix this by hand to some extent, and could write a jiffy to put it back into the multiple model format, but feel like there is a more elegant solution to this problem already embedded in Chimera.  Preferably one that will allow high throughput down the road.  Any advice would be appreciated!

Thanks,

Martin