Dear Minming, Dr. Ferrin forwarded your question to me. We recently added calculation of structure similarity measures to Chimera. This capability is not in versions 1.4/1.4.1, but only in 1.5 (recent daily builds, which are also available from our download page). In such recent builds, the "Match -> Align" tool (under Tools... Structure Comparison in the menu) calculates two measures that have been described in previous publications, the SDM and the Q- score. The SDM has been used to construct phylogenetic trees before. I don't know if the Q-score has been used that way. Documentation for Match->Align and details of these measures: <http://www.cgl.ucsf.edu/chimera/docs/ContributedSoftware/matchalign/matchali...
<http://www.cgl.ucsf.edu/chimera/docs/ContributedSoftware/matchalign/matchali...
However, there are some issues that may affect whether you decide to use Chimera to calculate these values for your structures: (1) You would need to run Match->Align for each pair. There is no option to do all pairwise comparisons on multiple structures. You would need to script looping through all pairs (with python or shell scripting). (2) There is no Chimera command for Match->Align. Thus Match->Align would also need to be run via python. I don't have scripting expertise, so someone else would have to provide more details if you decide to try this. (3) Before you run Match->Align, the structures need to be superimposed. We purposefully made Match->Align independent of the superposition method. There are several possible ways to superimpose structures in Chimera, <http://www.cgl.ucsf.edu/chimera/docs/UsersGuide/superposition.html> but probably the easiest in your case would be MatchMaker, which is also available as a Chimera command (matchmaker or mmaker, see that URL and links within). Usually the mmaker defaults work pretty well, for example Chimera commands: open 2mnr open 4enl mmaker #0 #1 (Then you could try running Match->Align interactively. The SDM, Q- score, etc. are reported in the Reply Log, opened with Favorites... Reply Log in the menu.) I hope this helps, Elaine ----- Elaine C. Meng, Ph.D. UCSF Computer Graphics Lab (Chimera team) and Babbitt Lab Department of Pharmaceutical Chemistry University of California, San Francisco P.S. we recommend sending Chimera questions to chimera-users@cgl.ucsf.edu (I CC'd that list here after removing the details from your original mail)
-------- Original Message -------- Subject: A question about your software(UCSF Chimera) Date: Tue, 23 Mar 2010 21:55:06 -0400
Dear Dr. Ferrin,
I am very interested in using your software UCSF Chimera, which have some commands that can do the job of aligning and measuring a distance between two protein structures.
I am a PhD student at Purdue University working on the evolution of protein structure with Professor Peter Waddell. I was hoping to use your structure alignment software to obtain a pairwise structure distance matrix. This matrix will then be used in phylogenetic analyses, comparing the trees and treeness of that structural distance compared to that produced by other published methods. [...] Could you please tell me if your structural alignment software can easily produce a complete pairwise distance matrix for a list of ~50 structures? I did not see a way to do this. If not, is there any easy way to modify it to do this? [...]
Best Regards and Sincerely, Minming Li
On Mar 30, 2010, at 9:04 PM, Elaine Meng wrote:
However, there are some issues that may affect whether you decide to use Chimera to calculate these values for your structures:
(1) You would need to run Match->Align for each pair. There is no option to do all pairwise comparisons on multiple structures. You would need to script looping through all pairs (with python or shell scripting).
(2) There is no Chimera command for Match->Align. Thus Match->Align would also need to be run via python. I don't have scripting expertise, so someone else would have to provide more details if you decide to try this.
Given that you said you wanted a complete pairwise matrix of ~50 structures, that would be ~1200 unique pairs, which is obviously way too many to do by hand. Currently Match->Align is awkward to script, even in Python, since all the important code is internal to the Match-
Align dialog and basically impossible to call externally. If you are sure that you would like to use Chimera for this computation I can reorganize the Match->Align code so that it could be called externally. I don't think the reorganization will be difficult. You would still need to resort to a Python script to make the Match->Align calls and loop through your models, so you would need some rudimentary familiarity with Python. Let me know.
--Eric Eric Pettersen UCSF Computer Graphics Lab http://www.cgl.ucsf.edu
Dear Dr. Pettersen and Dr. Meng, Thank you for your last email. I'm sorry that I am not so familiar with Python, so could you help me to use the Chimera to produce the distance matrix of the listed 45 pdbs below? Otherwise, is there any easy way to do this just by running the program instead of using the Match->Align by hand? Thank you so much! The list of the 45 pdbs: 1il2a,1asza,1kmmc,1adyb,1adjb,1b76b,1efwa,11asb,1j5wb,1bbwa, 1lylb,1e1oa,1pysa,1b8aa,1httb,1kmna,1qe0b,1g51a,1evlc,1evka, 1nj8d,1kmmb,1c0aa,1j5wa,1serb,1h4qa,1eqrc,1sera,1nj1a,1h4sb, 1atib,1h4vb,1b70a,1eqra,1kmnb,1adya,1adjc,1hc7a,1httc,1eqrb, 1g51b,1qe0a,1b76a,1atia,1efwb. Best regards! sincerely, Minming Li -------------------------- Minming Li PhD student, Interdisciplinary Life Science Program (PULSe Program), Purdue University, West Lafayette, IN. USA. Email: li362@purdue.edu Quoting Eric Pettersen <pett@cgl.ucsf.edu>:
On Mar 30, 2010, at 9:04 PM, Elaine Meng wrote:
However, there are some issues that may affect whether you decide to use Chimera to calculate these values for your structures:
(1) You would need to run Match->Align for each pair. There is no option to do all pairwise comparisons on multiple structures. You would need to script looping through all pairs (with python or shell scripting).
(2) There is no Chimera command for Match->Align. Thus Match->Align would also need to be run via python. I don't have scripting expertise, so someone else would have to provide more details if you decide to try this.
Given that you said you wanted a complete pairwise matrix of ~50 structures, that would be ~1200 unique pairs, which is obviously way too many to do by hand. Currently Match->Align is awkward to script, even in Python, since all the important code is internal to the Match-
Align dialog and basically impossible to call externally. If you are sure that you would like to use Chimera for this computation I can reorganize the Match->Align code so that it could be called externally. I don't think the reorganization will be difficult. You would still need to resort to a Python script to make the Match->Align calls and loop through your models, so you would need some rudimentary familiarity with Python. Let me know.
--Eric
Eric Pettersen UCSF Computer Graphics Lab http://www.cgl.ucsf.edu
Dear Dr. Li, I'm sorry, but that would not be feasible. It is not a "plug and chug" situation, where the exact series of tasks you need is already known. As mentioned previously, the structures would first need to be superimposed before Match->Align is used to calculate similarity scores. Although Chimera provides several ways to superimpose structures (Superimposing Structures), you would need to evaluate which method and parameters work on your structures of interest, using your own knowledge of this family of proteins. After superimposing, getting the scores, and building a phylogeny, if there were any discrepancies from what you were expecting, you would probably want to re-evaluate some of the superpositions and possibly recalculate them, again based on your own detailed knowledge of these proteins. We are willing to provide tools for your research, but actually performing this part of your study with Chimera would go far beyond that role. It sounds like you may need to take a simpler approach, such as using some other program that both superimposes proteins and provides a similarity score in the same step (as opposed to the decoupled approach in Chimera, which would also require python scripting knowledge). However, I'm not aware of any program that will do all pairwise comparisons automatically, so you would need to find a program that you can download for local use and then somehow (shell or other) script looping through all the pairs. Here is an informal list of superposition programs, some of which are available for download: Online Structure Alignment Resources If you recruit someone familiar with Python to your team, I'd be willing to make the scriptability improvements to Match->Align that I mentioned previously and provide guidance on the basic scripting outline you would need. --Eric Eric Pettersen UCSF Computer Graphics Lab http://www.cgl.ucsf.edu On Apr 3, 2010, at 2:09 PM, Minming Li wrote:
Dear Dr. Pettersen and Dr. Meng,
Thank you for your last email. I'm sorry that I am not so familiar with Python, so could you help me to use the Chimera to produce the distance matrix of the listed 45 pdbs below? Otherwise, is there any easy way to do this just by running the program instead of using the Match->Align by hand? Thank you so much!
The list of the 45 pdbs: 1il2a,1asza,1kmmc,1adyb,1adjb,1b76b,1efwa,11asb,1j5wb,1bbwa, 1lylb,1e1oa,1pysa,1b8aa,1httb,1kmna,1qe0b,1g51a,1evlc,1evka, 1nj8d,1kmmb,1c0aa,1j5wa,1serb,1h4qa,1eqrc,1sera,1nj1a,1h4sb, 1atib,1h4vb,1b70a,1eqra,1kmnb,1adya,1adjc,1hc7a,1httc,1eqrb, 1g51b,1qe0a,1b76a,1atia,1efwb.
Best regards!
sincerely, Minming Li
-------------------------- Minming Li PhD student, Interdisciplinary Life Science Program (PULSe Program), Purdue University, West Lafayette, IN. USA. Email: li362@purdue.edu
Quoting Eric Pettersen <pett@cgl.ucsf.edu>:
On Mar 30, 2010, at 9:04 PM, Elaine Meng wrote:
However, there are some issues that may affect whether you decide to use Chimera to calculate these values for your structures:
(1) You would need to run Match->Align for each pair. There is no option to do all pairwise comparisons on multiple structures. You would need to script looping through all pairs (with python or shell scripting).
(2) There is no Chimera command for Match->Align. Thus Match->Align would also need to be run via python. I don't have scripting expertise, so someone else would have to provide more details if you decide to try this.
Given that you said you wanted a complete pairwise matrix of ~50 structures, that would be ~1200 unique pairs, which is obviously way too many to do by hand. Currently Match->Align is awkward to script, even in Python, since all the important code is internal to the Match-
Align dialog and basically impossible to call externally. If you are sure that you would like to use Chimera for this computation I can reorganize the Match->Align code so that it could be called externally. I don't think the reorganization will be difficult. You would still need to resort to a Python script to make the Match- Align calls and loop through your models, so you would need some rudimentary familiarity with Python. Let me know.
--Eric
Eric Pettersen UCSF Computer Graphics Lab http://www.cgl.ucsf.edu
participants (3)
-
Elaine Meng -
Eric Pettersen -
Minming Li