On Fri, 2007-06-29 at 11:44 +0100, Kevin Cowtan wrote:
> OK, I can confirm the problem. The issue is that mama can make ccp4 maps
> which, while they technically obey the CCP4 map format, contain content
> which is ambiguous.
>
> In this case, your map extends from -29 to 53 on x, and yet the X-axis
> is only 44 grid points long - similar for y and z. So each grid point
> appears several times in the map, with possibly inconsistent values.
>
> Clipper (and therefore coot) use magic symmetry and cell repeat to allow
> you to visualise the unique contents of the structure anywhere in space,
> but as a result, they produce unpredictable results when the map header
> disagrees with the map contents.
>
> It is possible to fix the file by the following step:
>
> mapmask mapin my.msk mapout new.map << eof
> extend overlap
> xyzlim asu
> pad 0
> eof
>
> It would potentially be possible to include this step when reading a map
> in coot. (Clipper would read the map as a non-crystallographic map, and
> then convert it to a crystallographic map for coot). The problem is that
> sometimes this is *not* what you want to do. I'm not sure what the best
> way forward on this is - maybe a 'read mask' option which does exactly this.
How about a "Read USF Map" function/button?
Is there a clipper way to do what mapmask is doing for us? (I'd rather
not run a separate program with that creating a new map if we can munge
the USF map on the fly.) If not, we should fall back to running
mapmask, of course.
Paul.
|