Re: [ccp4bb]: DMMULTI or AMORE

[Kevin Cowtan <cowtan@ysbl.york.ac.uk>]

***  For details on how to be removed from this list visit the  ***
***          CCP4 home page http://www.ccp4.ac.uk         ***

Andrew Gulick wrote:
> My question is what is the best way to use the phase information
> from the first MAD data to phase the second native form.  It
> seems to me that I can use AMORE to generate tabling information
> from the hexagonal form F and Phi values.  Or I can use
> the multicrystal capabilities of DMMULTI to phase the orthorhombic
> form.  What I cannot figure out is if DMMULTI will only
> improve/compare phases from two data sets or if this program
> will take phases from one crystal form and transfer them to a
> second.  In other words, does it work if I have no phase
> information for the second crystal form or will it only improve
> the (previously obtained) phases from two different forms?

Yes, dmmulti will work just great for this task. On the first cycle, the
weight of the second form will be zero, since it has no phases, so the
calculation effectively just copies density across. The magnitude
information in the second form then ties these phases into the observed
data. On subsequent cycles, the averaging sends info both ways.

Improvmeent of the averaging operators is rather weaker in this case
than when phases are available in both forms, but if you run it for
enough cycles it ought to settle down. If it looks like it is
mis-behaving, you may need to play with the operator refinement steps,
or even split your run into two jobs.