Hi Jorge, In your version of phenix yes, you need a special format for the alignment files. I just changed it to allow a simple alignment file:

target sequence-of-target template sequence-of-template

which is also used by sculptor. This is available in a current nightly build. You can then just paste the alignment from hhpred into such a file. I hope that helps! -Tom T

...

...

Thanks, Tom.

Yes, phenix.mr_model_preparation fails just like phenix.mr_rosetta:

" Traceback (most recent call last): File "/usr/local/phenix-1.7.1-743/phaser/phaser/command_line/mr_model_preparation.py", line 9, in <mod ule> mr_model_preparation.run( args = sys.argv[1:], out = sys.stdout ) File "/usr/local/phenix-1.7.1-743/phaser/phaser/mr_model_preparation.py", line 255, in run proxy = proxy.DbfetchHTTP( database = "pdb", format = "pdb" ) File "/usr/local/phenix-1.7.1-743/phaser/phaser/proxy.py", line 108, in __init__ raise RuntimeError, "Cannot contact %s" % str( self ) RuntimeError: Cannot contact EBI DbFetch service "

I think I will have to go to phenix sculptor. This seems to require the alignment in ".aln (CLUSTAL format), .pir (PIR-format) and .ali (relaxed PIR-like format)" maybe I could prefer to keep the hhr alignment but I will have to check if there is an easy way to translate one format into the other. Apart from this, phenix.mr_rosetta, seems to require the alignment "MODELLER-style". Therefore, to keep the alignment as output by hhr (is this wise? Does this use multialignment and maight this be better that pairwise alignment?) I have to translate it to either aln/pir/ali (for phenix sculptor) and then to MODELLER-style for phenix.mr-rosetta. A suggestion to make this more straight forward?

J.

...

Hi Jorje,

Yes, you can prepare your model separately from mr_rosetta. You can do this (for example) with:

phenix.mr_model_preparation 3800765_mod_ed.hhr

If that fails with the same download problem, you can create an alignment file and use phenix.sculptor to edit your model.

Then as you suggest, you can use your edited model directly as a "search_model" and skip the .hhr files entirely.

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

...

...

Dear all,

I would like to use mr_rosetta with homolgy search. I followed the instructions at http://www.phenix-online.org/documentation/mr_rosetta.htm , generated the hhr and the fragment files, but at submitting the job I receive the message:

******************* ERROR ENDING *************** Cannot contact EBI DbFetch service

******************* ERROR ENDING ***************

after both loading the file 3800765_mod_ed.hhr and generating the log file mr_model_preparation.log.

Concerning this problem, I saw http://www.phenix-online.org/documentation/faqs.htm#anch8 . But I am not behind a proxy and even disabling completely the firewall, the message still appears. Maybe this is more linux related, I use openSuSE 11.4. One workaround would be that I write a script to wget the pdb files (listed in .hhr) and select the due chains, putting these in a certain location and indicating in my script where they are. But I am a kind of confused, might

search_models=

do this. I am not sure I could use a hhr file together with it. Anyway, this might be a desirable solution for me yet I use to run my stuff in a laptop which goes around by.

Thanks,

Jorge

_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

---

phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb