[phenixbb] questions about ncs copy in autobuild

Thomas C. Terwilliger terwilliger at lanl.gov
Wed Jul 21 14:08:26 PDT 2010

Hi Jason,

I am guessing that your starting model is an MR solution and you have data
that is not very high-resolution or the MR solution is not that close to
the structure that you are trying to solve.  In this case these results
would indicate that autobuild has failed to create a useful map of your
structure and then it is unable to build a useful model, unable to
identify sequence register, resulting in a single set of chains not
assigned to sequence (and probably mostly incorrect).

So a key question is: if you take your starting model and simply calculate
an R factor (with or without refinement), is this R much less than 0.5? 
If not, this will be very difficult to rebuild, particularly if the
resolution is not high.

An alternative question: the starting map in autobuild
(cycle_best_2.mtz)...does it look like your starting model? Does it look
like a protein?  If not....this will be difficult.

Let me know if that doesn't help!
All the best,
Tom T

>> Dear All,
>> I am using phenix.autobuild .eff file to build a structure (version
>> 1.6.1).
>> For the input I have a pdb model file with 6 chains; a sequence file which
>> contains residues of a single chain (the sequence file has a few more
>> residues at N-terminal than the pdb model file); an mtz amplitude file; I
>> also specified ncs_copies=6 (I know there are 6 copies in the au).
>> What confuses me is the output pdb file only has one chain (huge
>> Rwork&Rfree, >0.5). Should there be 6 chains?
>> The other issue is that the output chain has fewer residues than my
>> sequence
>> file (both N&C-terminal). It also has fewer residues compared with the pdb
>> model file. Is this because I do not have enough resolution to resolve
>> those
>> missing residues? Thanks in advance.
>> Jason
>> U Pitt
>> _______________________________________________
>> phenixbb mailing list
>> phenixbb at phenix-online.org
>> http://phenix-online.org/mailman/listinfo/phenixbb

More information about the phenixbb mailing list