[phenixbb] autosol strategy

Thomas C. Terwilliger terwilliger at lanl.gov
Thu May 28 13:17:44 PDT 2009


Hi Maia,

Some keywords have changed between 1.4-3 and the current versions.  I
think you'll like the current ones much better.   To see all the current
keywords, say:

phenix.doc

and select "Automated Structure Solution using AutoSol" and  and then
click on "List of all AutoSol keywords"

The keywords for MAD now are entered as a little parameters file:

autosol {
  seq_file = seq.dat
  sites = 2
  atom_type = Se
  wavelength {
    data = peak.sca
    lambda = .9798
    f_prime = -8.0
    f_double_prime = 4.5
  }
  wavelength {
    data = inf.sca
    lambda = .9792
    f_prime = -9.0
    f_double_prime = 1.5
  }

You can do that for SAD too, or just say

phenix.autosol data=peak.sca ha_type=Br sites=6 seq_file=seq.dat

If you have a heavy atom like Br where you really don't know how many
sites there are, you can just guess.  For SAD phasing, Phaser will put in
however many sites it can find whatever you say. For MAD phasing, SOLVE
will put in just the number of sites you ask for.  I would guess a number
about 1/20 the number of residues in the protein as a starting point.

All the best,
Tom T

 >> Hi Tom,
>>
>> how do I determine the number of heavy sites? I gave 2 as in your
>> example, but it can be more. Should I try several different numbers?
>>
>> Maia
>>
>> Maia Cherney wrote:
>>> Thanks Tom, Miguel,
>>> for your help.
>>>
>>> I would like you to know that the older version of phenix (1.4-3) works,
>>> but the newer versions 1.4-4 and 1.4-58 don't recognize some key words.
>>> The error message is
>>>
>>> Sorry, unknown file or keyword: peak.data=peak1_p6422.sca
>>>
>>>
>>> Maia
>>>
>>> Miguel Ortiz Lombardia wrote:
>>>
>>>> Hi Maia,
>>>>
>>>> If your phases are good enough I think resolve should be able to build
>>>> at least a good partial model. Otherwise, you can build it manually...
>>>> Your resolution isn't great, but you may be able to place some
>>>> secondary structure elements (ARP/wARP 'Quick fold' can also come to
>>>> help here).
>>>>
>>>> Perhaps another option is to solve your native dataset by molecular
>>>> replacement using the best map you have from the MAD dataset, then you
>>>> wouldn't need to build a model in the low resolution MAD data. Molrep
>>>> and AMoRe can do that, probably Phaser can as well.
>>>>
>>>> Good luck,
>>>>
>>>>
>>>> Miguel
>>>>
>>>> Le 28 mai 09 à 06:12, Maia Cherney a écrit :
>>>>
>>>>
>>>>> Hi Tom,
>>>>> thank you for your reply. How can I build a model at this resolution
>>>>> 3.3A (anomalous signal 4.2A). Actually, that was my major question. Is
>>>>> it possible to build a model at this resolution? Or how else I can use
>>>>> the phases from the low resolution solution (if I get it) with the
>>>>> high
>>>>> resolution native data?
>>>>>
>>>>> I did not see such a combination in the examples.
>>>>>
>>>>> Maia
>>>>>
>>>>> Thomas C. Terwilliger wrote:
>>>>>
>>>>>> Hi Maia,
>>>>>>
>>>>>> As your datasets are in 2 different space groups, you pretty much
>>>>>> have to
>>>>>> solve the structure in one space group, then transfer information
>>>>>> about
>>>>>> the structure (model) to the other space group.
>>>>>>
>>>>>> So...you want to run an autosol run with the MAD data (or as SAD,
>>>>>> sometimes that works better if you have a lot of decay) to solve the
>>>>>> structure...then you want to build a model if you can in that space
>>>>>> group.
>>>>>>
>>>>>> Then you have to figure out where that model goes in the high-res
>>>>>> crystal...you'll use phenix.automr  with your model from the low-res
>>>>>> crystal to do that.
>>>>>>
>>>>>> Then you can do cross-crystal averaging with phenix.multi_crystal
>>>>>> average
>>>>>> to improve the phases (optional). You can also just rebuild the
>>>>>> model in
>>>>>> the high-res crystal form and you may be done.
>>>>>>
>>>>>> Good luck!
>>>>>> All the best,
>>>>>> -Tom T
>>>>>>
>>>>>>
>>>>>>
>>>>>>>> Hi everyone,
>>>>>>>>
>>>>>>>> I have a native set to 2.3A resolution and mercury MAD data in a
>>>>>>>> different space group to 3.3A resolution (anomalous signal to
>>>>>>>> 4.2A).
>>>>>>>>
>>>>>>>> How I can use autosol with the data I have. There is no such
>>>>>>>> example.
>>>>>>>> (Native +MAD)? What is the best strategy in this case?
>>>>>>>>
>>>>>>>> Maia
>>>>>>>>
>>>>>>>>
>>>>>>>> _______________________________________________
>>>>>>>> phenixbb mailing list
>>>>>>>> phenixbb at phenix-online.org
>>>>>>>> http://www.phenix-online.org/mailman/listinfo/phenixbb
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>> _______________________________________________
>>>>>> phenixbb mailing list
>>>>>> phenixbb at phenix-online.org
>>>>>> http://www.phenix-online.org/mailman/listinfo/phenixbb
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>> _______________________________________________
>>>>> phenixbb mailing list
>>>>> phenixbb at phenix-online.org
>>>>> http://www.phenix-online.org/mailman/listinfo/phenixbb
>>>>>
>>>>> --
>>>>> This message has been scanned for viruses and
>>>>> dangerous content by MailScanner, and is
>>>>> believed to be clean.
>>>>>
>>>>>
>>>> --
>>>> Miguel Ortiz Lombardía
>>>> Architecture et Fonction des Macromolécules Biologiques
>>>> UMR6098 ( CNRS, U. de Provence, U. de la Méditerranée )
>>>> Case 932
>>>> 163 Avenue de Luminy
>>>> 13288 Marseille cedex 9
>>>> France
>>>> Tel : +33(0) 491 82 55 93
>>>> Fax: +33(0) 491 26 67 20
>>>> e-mail: miguel.ortiz-lombardia at afmb.univ-mrs.fr
>>>> Web: http://www.pangea.org/mol/spip.php?rubrique2
>>>>
>>>> ------------------------------------------------------------------------
>>>>
>>>> _______________________________________________
>>>> phenixbb mailing list
>>>> phenixbb at phenix-online.org
>>>> http://www.phenix-online.org/mailman/listinfo/phenixbb
>>>>
>>>>
>>>
>>> _______________________________________________
>>> phenixbb mailing list
>>> phenixbb at phenix-online.org
>>> http://www.phenix-online.org/mailman/listinfo/phenixbb
>>>
>>>
>>
>> _______________________________________________
>> phenixbb mailing list
>> phenixbb at phenix-online.org
>> http://www.phenix-online.org/mailman/listinfo/phenixbb
>>




More information about the phenixbb mailing list