Extracting chunks of data based on variables stored in another file (Perl?)

mchriste · 03-05-2009, 01:05 PM

I need to extract chunks of information from a file based on two variables that are stored within another file.

I have attached an example "data" file (alldocked.pdbqt) and "variables" file (roots.pdb).

Important notes:
remove the .txt extension first (required by the forum)

I had to heavily snip the data file. A real data file can easily be 10MB large.

The format for each record I want to extract is the following:

"Data" file: alldocked.pdbqt

Code:

<filename>:DOCKED: MODEL        <model>
<filename>:DOCKED: USER (...)
(...)
<filename>:DOCKED: (all kinds of stuff, many lines)
(...)
<filename>:DOCKED: ENDMDL
<filename>:DOCKED: MODEL        <model>
<filename>:DOCKED: USER (...)
(...)
<filename>:DOCKED: (all kinds of stuff, many lines)
(...)
<filename>:DOCKED: ENDMDL
(etc)

The <filename> and <model> variables are stored in a separate file called roots.pdb.
The format of roots.pdb is for each line:

Code:

ATOM      1    O UNK L   0       <coord1>   <coord2>  <coord3> 1.00  <energy>           O <filename> <model>

Characteristics of the data:

<filename> corresponds to field #13, <model> to field #14 in each line.
Both <filename> and <model> vary INDEPENDANTLY for each record.
<filename> can be absolutely random but it will always end with ".dlg".
There may be up to 99 different <filename> in the alldocked.pdbqt file.
<model> is always an integer and can range from 1 to 999.
The "data" file is always called alldocked.pdbqt, the "variables" file roots.pdb

WHAT I WOULD LIKE TO GET:
I need a file that contains the records corresponding to the first x (say, 20) lines of roots.pdb.
I also need to add a line containing the <filename> between each record.
(Yes I know that <filename> is included on each line, but later I will strip this).

So, the output format I desire would be: (output.pdbqt)

Code:

<filename>:DOCKED: REMARK <filename>
<filename>:DOCKED: MODEL        <model>
<filename>:DOCKED: USER (...)
(...)
<filename>:DOCKED: (all kinds of stuff, many lines)
(...)
<filename>:DOCKED: ENDMDL
<filename>:DOCKED: REMARK <filename>
<filename>:DOCKED: MODEL        <model>
<filename>:DOCKED: USER (...)
(...)
<filename>:DOCKED: (all kinds of stuff, many lines)
(...)
<filename>:DOCKED: ENDMDL
(...)

Enough with the abstract descriptions, here's a real example:

The example roots.pdb file I attached starts with the following lines:

Code:

ATOM      1    O UNK L   0       9.200   5.778  -4.800 1.00  -10.02           O SOS_charged1.dlg  13 
ATOM      2    O UNK L   0      13.230   3.129  -6.038 1.00  -7.94           O SOS_charged2.dlg  20 
ATOM      3    O UNK L   0      11.295   0.656  -6.503 1.00  -7.91           O SOS_charged1.dlg   7

So, the output file I would like should start with:

Code:

SOS_charged1.dlg:DOCKED: REMARK SOS_charged1.dlg
SOS_charged1.dlg:DOCKED: MODEL        13
(...)
SOS_charged1.dlg:DOCKED: (all data in alldocked.dlg that corresponds to model SOS_charged1.dlg 13)
(...)
SOS_charged1.dlg:DOCKED: ENDMDL
SOS_charged2.dlg:DOCKED: REMARK SOS_charged2.dlg
SOS_charged2.dlg:DOCKED: MODEL        20
SOS_charged2.dlg:DOCKED: USER (...)
(...)
SOS_charged2.dlg:DOCKED: (all data in alldocked.dlg that corresponds to model SOS_charged2.dlg 20)
(...)
SOS_charged2.dlg:DOCKED: ENDMDL
SOS_charged2.dlg:DOCKED: REMARK SOS_charged2.dlg
SOS_charged1.dlg:DOCKED: MODEL        7
SOS_charged1.dlg:DOCKED: USER (...)
(...)
SOS_charged1.dlg:DOCKED: (all data in alldocked.dlg that corresponds to model SOS_charged1.dlg 7)
(...)
SOS_charged1.dlg:DOCKED: ENDMDL

Who can help me? I'm trying to analyze protein-ligand docking results in a way that is not provided by the software I'm using - but I'm a biochemist, not a programmer!

mchriste · 03-05-2009, 01:08 PM

Addon: I figure that a PERL script would be the way to go.
Something like (pseudocode):

Code:

#!/usr/bin/perl
print "How many conformations to extract? "
chomp ($xmax = <>);

$FILENAME=`awk '{print $13}' roots.pdb`;
$MODEL=`awk '{print $14}' roots.pdb`;

$START = "concatenate("$FILENAME,:DOCKED: MODEL        ,$MODEL")";
$END = "ENDMDL";

undef $/;
$_ = <STDIN>;

$line=1

print concatenate("$FILENAME,:DOCKED: REMARK ,$FILENAME");

if (/($START.*?$END)/s) {
print $1;
}

(append to output.pdbqt);

$line++;

(repeat the above until $line = $max)

mchriste · 03-12-2009, 12:44 PM

bump - did I confuse everybody with my problem?
I thought my second post clarified what I need...