Member Site › Forums › Rosetta 3 › Rosetta 3 – Applications › the extracted pdb file
- This topic has 12 replies, 3 voices, and was last updated 11 years, 7 months ago by
Anonymous.
-
AuthorPosts
-
-
January 5, 2013 at 1:55 am #1494
Anonymous
Hi,
I used the rosetta3.4_membrane_homology_modeling to build a pdb file. but when I extract them from the silent file, my target pdb didn’t looks like a normal protein, even like a jumble of segments.Did I make a mistake?
the silent file was created by this command line: ~/rosetta3.4/rosetta_source/bin/minirosetta.linuxgccrelease @flags, and the flag file was attached.
to extract the pdb file from silent file by: ~/rosetta3.4/rosetta_source/bin/extract_pdbs.linuxgccrelease @convert.flags,and the convert.flag was also attached.
the created pdb file was attached in the third attachment. please help me~~~
Thanks!!
xiaofeng
-
January 7, 2013 at 4:02 pm #8279
Anonymous
You have a silent file type mismatch. I’m surprised it gives you this garbage output instead of just crashing.
In your flags2: -out:file:silent_struct_type binary
In your convert_flags2: -in::file::silent_struct_type proteinChange “protein” to “binary” on the latter and see if it works better on the existing domain1.out.
-
January 8, 2013 at 12:47 am #8287
Anonymous
Dear smlewis,
I have changed the “protein” into “binary” on the latter command line, but the result did none change. They are just the same as before.
Should I do some other work on the extracted pdb file?
Thanks!
Xiaofeng
-
January 8, 2013 at 1:12 am #8288
Anonymous
OK, two things to try:
1) Re-run your first command (minirosetta), but without the commands “-out:file:silent_struct_type binary” and “-out:file:silent domain1.out”, replacing them with “-out:pdb” (I think the latter is unnecessary but it won’t hurt). Lower nstruct to something that will run quickly for a test, 100 or less is ok. Do these PDBs come out okay?
2) Post your log file from the second (extract_pdbs) job as an attachment so I can see if I spot a problem.
-
January 9, 2013 at 1:36 am #8299
Anonymous
A) I meant for you to REMOVE the flags “out:file:silent_struct_type” and “out:file:silent” completely, replacing them with the flag -out:pdb.
Look at the first ten lines of the “domain1.out” you produced using “silent_struct_type binary” and then at the first ten lines of the “domain1.out” you produced using silent_struct_type “pdb” (those are separate files in separate folders, right? Otherwise this will certainly not work). You say you had no change in your results, but those files should look *very* different from each other. The first should have lots of lines that look like gibberish, like:
L0isJBVO0MEksdeiQnv/MB1isTEU46LiQLyWNBpw1BEkuJ7hQhrnOBB/peDkVOAiQJbnSBdSMaEEtITiQv0NWBdSMiEUzMBiQUN+aBJGEuE025JiQ0iMaB1is9EUO0UiQVjXfBpvfmEUwKHiQb8SLBVjXMEkrHtiQQ1YGBpa8UE0eUiiQ6mEIB5kY8DkAreiQ1NJLBVNehEERLEiQEtISBxJxmEEnEfiQtIbUB5QLMEU3kaiQsHlXBdTiVEUxg2hQIwqTBR2OtEUjX4hQ msd_output_10_LHp_1_0092
The second should have lots of lines that look like the PDB file format.
-
January 9, 2013 at 4:02 am #8298
Anonymous
Dear smlewis,
The command I changed it into:
-out:file:fullatom
-out:file:silent_struct_type pdb
-out:file:silent domain1.out
and the command in extraced script is : -in::file::silent_struct_type pdbAm I right? If these commands are right, the result has none change.
I wonder the correct output pdb file should be. could I get a copy as an example from you?
Thank you very much for your patience~
Best regards!
-
January 9, 2013 at 8:29 am #8300
Anonymous
Yes, smlewis , you are right. There are the differences as you mentioned.
What I say “the result has none change” is for the extraced pdb file. When I displayed the pdb file in pymol, they were the same, they did none change with the chege of the command line.
later I remove the flags “out:file:silent_struct_type” and “out:file:silent” completely, replacing them with the flag -out:pdb. The result also like before.
the log file was attached, please check, Thanks!
-
January 10, 2013 at 3:37 am #8302
Anonymous
Maybe these error were given by the loop seting.
In the loop file, I set it according to the OCTOPUS output file,
for example:
OCTOPUS predicted topology:
iMMMMMMMMMMMMMMMMMMMMMoooooooooooMMMMMMMMMMMMMMMMMMMMMiiiiii
iiiiiiiMMMMMMMMMMMMMMMMMMMMMoooooooooooooooooooooooooooooooo
ooooMMMMMMMMMMMMMMMMMMMMMiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiii
iiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiMMMMMMMMMMMMM
MMMMMMMMoooooooooMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMithe loop file:
LOOP 23 33 0 0.0 X
LOOP 55 67 0 0.0 X
LOOP 89 124 0 0.0 X
LOOP 146 227 0 0.0 X
LOOP 249 257 0 0.0 XBut your tutorial mentioned that this loop file should be build according to zone.
if the zone file like this:zone 1-55 : 1-55
zone 61-82 : 56-77
zone 86-151 : 78-143
zone 227-289 : 144-206is the loop file should like this?
LOOP 56 60 0 0.0 X
LOOP 152 226 0 0.0 XThere is a question, dose the segment don’t homology with the template pdb, it is sure to form a loop?
-
January 10, 2013 at 5:26 am #8303
Anonymous
Sorry, rmoretti, I think I did not understand rosetta_membrane_homology_modeling completely. Now I realize the reason for these errors. I built the loop file according to zone, the result pdb looks like an ordinary pdb file.
I Sincerely Apologize For All The Trouble I’ve Caused.
Thanks and best regards!
Xiaofeng
-
January 10, 2013 at 5:52 am #8304
Anonymous
Dear rmoretti,
I have another question about the loop file setting.
if the zone file like this:
zone 1-55 : 1-55
zone 61-82 : 56-77
zone 86-151 : 78-143
zone 227-289 : 144-206is the loop file should like this?
LOOP 56 60 0 0.0 X
LOOP 152 226 0 0.0 X
LOOP 83 85 0 0.0 XAs far as I know the fragments in the fragment_database formed by rosetta online server has only two length, 3 and 9. I wonedr whether the length of dismatch zone in the zone file is less than two, it can be delete? or all of the dismatch zone must be listed in the loop file?
Best Regards!
xiaofeng
-
January 14, 2013 at 12:35 am #8306
Anonymous
Dear rmoretti,
Thank you very much for your help, now I know the inherent meaning of loop modeling. The result appeared noemal.
Best Regards!
Xiaofeng
-
January 9, 2013 at 7:10 pm #8301
Anonymous
If using the -out:pdb flag gives you a similar PDB to the one your get with -out:file:silent_struct_type binary -> extract_pdbs, then it’s not the extraction process that’s giving you the errors, it’s how you’re running the the initial homology modeling. After talking with others in the lab, it’s likely that there’s issues with your template and your alignments, though it’s hard to say at this point.
Keeping the -out:pdb and a low -nstruct, I’d recommend playing around with the modeling run, and seeing what’s causing the disjointed output. First thing to check is the format of all of the input files. Do the input PDBs look reasonable? Are the other files formatted appropriately? (Keep in mind that spacing and missing/extra lines sometimes matters.) Is it a particular flag that’s causing issues? (Try removing flags that start optional/additional protocols. Or try altering which protocol you use – e.g. try kinematic loop remodeling instead of CCD.) Another thing to try is running things as a normal, non-membrane modeling job. Do you see the same errors then?
If you think the input files look reasonable, please post them, particularly the out.pdb native/loop input, the domain1.pdb -s starting structure, as well as the loop file and the span file. Troubleshooting becomes easier when we have enough information to replicate the issue on our own machines.
-
January 10, 2013 at 9:44 pm #8305
Anonymous
Loops shorter than 3 aa are perfectly fine. In addition to the 3 and 9-mer fragments, Rosetta also has a database of “1-mer fragments” (basically, amino acid backbone conformation propensities), that don’t need to be specified on the command line as they’re protein independent. In addition, Rosetta will sometimes trim back the protein to expand the loops some, if necessary.
You should include all the loops you want to have Rosetta close (e.g. anywhere there would be a chain break in your model) in the loops file, even if they’re only a few amino acids long. Note that Rosetta might have a harder time closing short loops than longer loops, just because of an inability to find the appropriate fragment closure. In that case it might make sense to increase the size of the loop (and correspondingly decrease the size of the matched region) to permit more backbone flexibility. But that’s only if the normal way of doing things doesn’t give you the results you want.
-
-
AuthorPosts
- You must be logged in to reply to this topic.