[phenixbb] Issue with phenix.superpose_maps

Nathaniel Echols nechols at lbl.gov
Fri Dec 3 07:18:08 PST 2010


On Fri, Dec 3, 2010 at 5:33 AM, Stephen Graham <scg34 at cam.ac.uk> wrote:
> I'm trying to use phenix.superpose_maps through the GUI to get
> superposed maps covering the same region of NCS-related chains A and B
> for a figure.  I have used PDB Tools to split the model into two files
> with only chain A or chain B.  However, when I run superpose_maps
> specifying the two individual chains for the input PDB files but the
> same mtz file for the map coefficients I only get one output map (I
> think it overwrites the output file...).  Selected input from the
> generated .eff file is as follows:
>
>    saving CCP4 map to Native_refine_6_map_coeffs_superposed.ccp4
>    saving CCP4 map to Native_refine_6_map_coeffs_superposed.ccp4
>
> A bug?

Definitely.  Will fix this, but you could work around it for now a
second copy of the MTZ file for now, and use that as the second input.

> P.S.  Incidentally, is there any reason you write both maps in a P1
> box rather than putting map 2 on top of map 1 in the position of the
> first (reference) model?  The latter would be more convenient as you
> can then re-use your carefully chosen pymol viewing matrices for all
> figures...

When Pavel and I were working on the map superpositioning we kept
running into the problem of unit cell periodicity screwing up the
result.  Transforming everything into a P1 box with a large buffer
around the molecules was the easiest way around the problem, but I
agree that it's not ideal (it was written for a different project that
needed results quickly).  I think the CCP4 map format should allow us
to do what you're asking, but I'm not sure if the maps API in CCTBX
will - need to check this with Ralf.

-Nat



More information about the phenixbb mailing list