[Chimera-users] mmaker

[Eric Pettersen]

On May 2, 2005, at 7:25 PM, Bryan W. Lepore wrote:

> hi Eric
>
> On Mon, 2 May 2005, Eric Pettersen wrote:
>> 	mmaker is a command for use in Chimera's "internal" command
>> line...
>
> ok,...
>
>> It can be included inside scripts that are run from the Unix command
>> line though.
>
> so like this?:
>
> #!/bin/bash
> chimera --nogui << eof
> mmaker
> file1.pdb
> file2.pdb
> eof
>
> this fails, but there's no output, hence the email.

Chimera can handle two kinds of script files:  those containing Chimera 
command-line commands (file name ending in .com or .cmd or preceded 
with cmd: or com: on the Unix Chimera startup command-line) and Python 
files (.py ending or py: prefix).

So if you had a file named "domatch" that contained:

	open file1.pdb
	open file2.pdb
	mmaker #0 #1 pair bb
	write #1 transformed.pdb

and executed chimera as "chimera --nogui cmd:domatch" then file2 would 
be superimposed on file1 and the transformed coordinates of file2 would 
be written out to transformed.pdb

>
>> 	If you briefly outlined what you want to do, I could provide better
>> advice about your best approach.
>
> take a list of .ent files and generate a superposition-based sequence
> alignment.  i can do this ok with the gui.

There are no Chimera commands that write sequence alignments, so this 
can only be done using Chimera's Python layer.

I've appended a Python script that does the following:

	1) reads a file named "structList" that has lines consisting of 3 file 
names:  2 PDB files and an output alignment file

	2) uses MatchMaker to superimpose the structures

	3) uses the "iterate match" capability of Multalign Viewer to improve 
the superposition by pruning off long residue pairings and rematching

	4) creates a superposition-based sequence alignment using Match->Align

	5) uses Multalign Viewer again to write out the alignment

Now, the script can't be run in nogui mode since Multalign Viewer is 
gui-only at this point.  Nonetheless, it does run without user 
intervention, so you can just leave it running on a display you aren't 
using.  There are comments in the script about what the code is doing 
and what various options are.  Let me know if you have questions.

I should also note that the script is using Needleman-Wunsch to do the 
initial sequence alignment.  If your structures have low sequence 
identity, you might want to switch to Smith-Waterman instead.  I am 
working on an improvement for MatchMaker that will make it much better 
at matching low sequence identity structures (it will consider 
secondary structure in addition to residue type).  I could send you the 
new version when it's ready in a few days if the current MatchMaker has 
trouble on your data.

--Eric


-------------- next part --------------
A non-text attachment was scrubbed...
Name: structAlign.py
Type: application/octet-stream
Size: 2434 bytes
Desc: not available
URL: <http://plato.cgl.ucsf.edu/pipermail/chimera-users/attachments/20050503/95878048/attachment.obj>