# How to extract specific pdb from silent files

6 posts / 0 new
How to extract specific pdb from silent files
#1

There is silent files which includes 160,000 models.

I am going to use extract_pdbs.mpi.linuxgccreleasee command for extracting low score top 100 or top 10.

What extra commands or options do I insert in?

Post Situation:
Tue, 2013-05-14 05:29
eunwook

You should be able to specify which structures to extract with the flag -in:file:tags. Simply pass the one or more tag for the structures you want to extract.

Tue, 2013-05-14 10:51
rmoretti

What is mean .tags files?

Is there any exprain for .tags fies in manual?

Wed, 2013-05-15 23:05
eunwook

It's not a file - it's a commandline parameter. The tags are the names that each structure is given in the silent file.

Say you have a silent file, and want to extract the lowest 5 energy structures. You can extract the score lines from the file like so:

grep SCORE silent.out > silent_scores.tab

This will give you a file with lines like the following:

SCORE: score fa_atr fa_rep fa_sol fa_intra_rep ... time description
SCORE: -3359.805 -1276.795 67.992 429.112 0.014 ... 160.000 3QF7X_0000015_0
SCORE: -3363.060 -1275.217 67.715 428.002 0.014 ... 126.000 3QF7X_0000021_0
...

This gives the scores for each of the structures in the silent file, along with the tag (under the description column).

So to extract the 5 lowest scoring structures from the silent file I would do something like

extract_pdbs.linuxgccrelease -database /path/to/database/ -in:file:silent silent.out -in:file:tags 3QF7X_0000025_0 3QF7X_0000032_0 3QF7X_0000413_0 3QF7X_0000175_0 3QF7X_0000378_0

(Add extra flags to the command line as appropriate)

Thu, 2013-05-16 11:32
rmoretti

Thank you rmoretti.
Is there a simple way to extract 1000 lowest-score pdb files among the 10000 models, for example?
If I use this way, then I have to identify 1000 tags and paste them in the command line or a file.

Thu, 2013-05-23 18:55
eunwook

Hi eunwook !
rmoretti already told you how to create a file containing the pdbs of your interest.
you can also do it if you have a score file using simple UNIX commands.
like $cat score.sc | sort -nrk col_number | tail -1000 | awk '{print$NF}' > pdb_list

Now just provide this file in the tag like : -in:file:tags cat pdb_list
Here is the manual: https://www.rosettacommons.org/manuals/archive/rosetta3.1_user_guide/app...

Fri, 2013-05-24 00:49