Re: [ccp4bb] Can't choose full H-M symbol for conventional cell ?

Protein crystallography Programs for crystallography X-ray detectors
Basic tutorials:
Protein Peptide

CCP4bb navigation

CCP4bb <-- 2007 <-- March 2007 <-- 18 March 2007

Previous message:
Subject: RA position in Genentech X-ray group
From: Sarah Hymowitz hymowitz {- dot -} sarah {- at -} GENE {- dot -} COM
Date: 2007-03-16

Next message:
Subject: Re: Can't choose full H-M symbol for conventional cell ?
From: Petrus H Zwart PHZwart {- at -} LBL {- dot -} GOV
Date: 2007-03-18

Subject: Re: Can't choose full H-M symbol for conventional cell ?
From: Jim Pflugrath jim {- dot -} pflugrath {- at -} RIGAKU {- dot -} COM
Date: 2007-03-18

Hi Ian,

I am happy to read that your users trust d*TREK so much that they simply
click the buttons and away they go. The indexing program within d*TREK
now outputs the following text:

INFO - The above indexing solution is ONLY a hypothesis.
One must confirm the hypothesis by examining symmetry
of observed intensities as well as possible systematic
absences.
One does this by collecting data, processing the data
and using dtcell and dtscaleaverage.

But my guess is that folks don't read the output in this click-happy
world. Indeed, as I have told many folks, the indexing step is the first
step in a crystal structure determination where one might need to know
some crystallography before proceeding. One can clone, express, purify,
..., grow crystals, harvest, flash-cool them, take diffraction images --
all without even knowing that there was someone named Bragg who had a Law.

There are even scripts or pipelines where users put in multiple
spacegroups and let the algorithm decide which one works to solve the
structure.

So I view this as one of education. It is probably easier to teach your
colleagues how to transform cell axes than it is to teach them how to
interpret or even write Hermann-Mauguin symbols.

d*TREK can help you with the transformation. One can transform cell axes
and reindex diffraction datasets trivially with dtcell.

One can even force dtindex to only look for a=70, b=90, c=60 and
spacegroup 18 as P2(1)2(1)2, if you like. So I cannot imagine that your
problems cannot be solved with existing solutions. First, confirm with
the user that the spacegroup number is 18 and that they want the 60
Angstrom axis to not have systematic absences. Then you can use
dtindex ... -cell 70 90 60 90 90 90 -known -spacegroup 18 ...
This can be used in scripts or in the GUI. For more help on dtcell, see
"dtcell -h", but a possible option is
dtcell ... -transform bca ...

Cheers,

Jim

> Subject: [ccp4bb] Can't choose full H-M symbol for conventional cell ?
>
>
> We're having some problems essentially because AFAICS Mosflm (and also
> D*Trek) don't recognise the 'full' Hermann-Mauguin symbols for space
> groups.
>
> For example for space group no 18 with the 'short' H-M symbol 'P 2/1 2/1
> 2' in the setting with 'full' H-M symbol 'P 2 2/1 2/1' the programs
> choose the correct conventional cell as say a=60, b=70, c=90 (according
> to the convention a<=b<=c). But then unless I've missed something
>...

CCP4bb navigation

CCP4bb <-- 2007 <-- March 2007 <-- 18 March 2007

Previous message:
Subject: RA position in Genentech X-ray group
From: Sarah Hymowitz hymowitz {- dot -} sarah {- at -} GENE {- dot -} COM
Date: 2007-03-16

Next message:
Subject: Re: Can't choose full H-M symbol for conventional cell ?
From: Petrus H Zwart PHZwart {- at -} LBL {- dot -} GOV
Date: 2007-03-18