I have a question about the suggestion given in the documentations of ab initio modeling: "to increase your chance of sampling the correct topology, a diverse set of homologous sequences, preferably with sequence changes that may have a greater impact on sampling like deletions and differences in conserved positions, may also be run since a homologue may converge towards the native structure with significantly less sampling." After generating a number of decoys of these homologous sequences, it is recommended to map these decoys to the target sequence.
Does anyone have experiences in doing this?
I have tried comparative modeling method with the silent files that containing many decoys of homologs as template_silent files. But it does not work. Probably, I used wrong options. I paste the options below. If there is something wrong, please give me some suggestions:
-loops:frag_sizes 9 3 1
-loops:frag_files inputs/aatargA09_05.200_v1_3 inputs/aatargA03_05.200_v1_3 none
It may be helpful to clarify what you mean by "does not work". (Does it exit with a error? If so, what error? Does it not give output structures? Does it give results, but not what you expected? If so, what did you get, and what did you expect?)
The job starts with reading sequence, template_silent file, fragment files and then stops with the complain like this:
protocols.jd2.JobDistributor: no more batches to process...
protocols.jd2.JobDistributor: 0 jobs considered, 0 jobs attempted in 0 seconds
protocols.jd2.JobDistributor: no jobs were attempted, did you forget to pass -overwrite?
I also notice a warning after it read in the template silent file:
protocols.jd2.ThreadingJobInputter: Warning: no template pdb provided for alignment tempA
The "tempA" is supposed to be the template. I have put the sequence of the template in the alignment file.
It seems that the alignment file is not recognized. Is this due to that the name of the template sequence is different from structures in the template_silent file (the tag of these structures in the silent file).
That's probably it. From the (almost) complete listing of options (it's not easy to find, and in general may be more confusing than helpful - http://www.rosettacommons.org/manuals/archive/rosetta3.3_user_guide/opti... ):
-template_silent <File> -- input templates for comparative modeling -- tag needs to fit alignment id
Change the name in the alignment to match that of the silent file tag, and see if that fixes your problem.
I have changed the name in the alignment to match that of the silent file tag, but it still complains that "no template pdb provided for alignment tag_name".
Does anyone know how to map decoys of homologs back to the target sequence? What is the general procedure for this kind de novo prediction? The online manual only gives the suggestion that do de novo predictions for some homologous sequences and map back to the target sequence.
Thank you a lot!
I have checked the code of minirosetta and found that the function only compares five digital numbers starting from the third character of the tag with the name that of the alignment file. So I changed the name of the template sequence in the alignment file and it works.
But there still something making me uncomfortable. The tags of all decoys in the output silent file are the same because the program output tags using the same name of target sequence as that used in the alignment file.