Reading PDB files

[esswired]

I am a very non-technical user, having never even opened a command window before I downloaded perl in an attempt to read a PDB file. I would love to be able to pull the PDB file into Excel (or Access which I think I could learn to use). Am I at the right place? Is there an easier way I should know about? Respectfully . . .

[scain]

Where to start?

First, I would really suggest the book Learning Perl, authored by our very own merlyn.

Next, get yourself to CPAN, the Comprehensive Perl Archive Network. There are literally thousands of useful Perl modules for doing all sorts of things, like reading special file formats and working with other programs. Since you are probably using Perl on Windows, and therefore you probably have Activestate Perl, you will want to learn how to use PPM to install those modules on your machine.

Finally, if by PDB file you are referring to a Protein DataBase Bank file, you will want to check out BioPerl, which has many useful bioinformatics functions, including reading special file formats. You might also want to consider gettting Beginning Perl for Bioinformatics, which is a pretty good reference for a scientist who wants to learn bioinformatics.

Good luck, and feel free to come back with more specific questons when you've started writing code.

Scott
Project coordinator of the Generic Model Organism Database Project

[Anonymous Monk]

Check out:

PDB Perl Module

This provides a mechanism for Perl to read write PDB (Protein Data Bank) files. I assumed this is the file type you mean rather than a palm pilot database. You could use this module to create a comma separated data file to import into Excel. You might also want to check out bioperl

Arun

[PodMaster]

Palm::PDB just in case (this also)

MJD says you can't just make shit up and expect the computer to know what you mean, retardo! ** The Third rule of perl club is a statement of fact: pod is sexy.

[esswired]

I have a Palm PDB file & want to read the whole thing & want to change it into a CSV file. Any ideas?

[gt8073a]

I am a very non-technical user

Bioperl may be way more complicated than you need to get started( don't get me wrong, it is fabulous, just not simple to follow )

The easiest approach is to read each line of the file, and matching the words/characters up to the first colon. How you store the data really depends on what data you want and what you intend to do with it all( again, this can get extremely tricky ).

Here is an extremely simple script that reads all sequences for an entry, blast them, and writes the information to a file. If you refer to your Camel book, this script should be easy enough to figure out and extend. Remember, though, Excel can read csv( comma seperated value ) files, and you will need to read up on DBI if you want to throw data into Access.

#!/usr/bin/perl -w

use strict;        ## watch typos and errors
$|++;            ## output

my $begin        = time;    ## just seeing how long we take

## change these
my $root        = 'C:\Windows\';    ## where the base of our script is
my $source        = "$root/PDBFind.txt";    ## do NOT use the full PDB
+Find.TXT file, I mean it!
my $out            = "$root/out.txt";    ## file to save blast results
+ to

my $blast_exe        = 'C:\Blast\bl2seq.exe';    ## where our blast pr
+ogram is
my $options        = '-G 11 -E 1 -W 3 -X 50 -F F -p blastp';    ## our
+ blast options

my $min_size    = 10;    ## don't blast shorty sequences, this isn't a
+ protein anyway

## prolly don't change these
my $blast_i        = "$root/i.$$";    ## temp files to hold sequences
my $blast_j        = "$root/j.$$";

## quick recap so we know what we are doing and where we are
print <<"ENDOFPRINT";
SCRIPT    : $0
IN    : '$source'
OUT    : '$out'
BLAST    : '$blast_exe $options -i "$blast_i" -j "$blast_j"'
ENDOFPRINT

## lines in PDBFind I care about
my $DELIMITER    = '//';
my $ID        = 'ID           : ';
my $CHAIN    = 'Chain        : ';
my $SEQUENCE    = ' Sequence    : ';
my $COMPOUND    = 'Compound     : ';
my $SOURCE    = 'Source       : ';
my $HEADER    = 'Header       : ';

## counters
my $pdb_count    = 0;
my $chain_count    = 0;
my $blast_count    = 0;
my $wild_count    = 0;
my $stub_count    = 0;
my $dna_count    = 0;

## variables relevant to data I am looking for
my ( $id, $chain );
my ( @chain_seq_pairs, @chain_wild, @chain_stubby, @chain_dna );

## open files
open( SOURCE, "<$source" )
    or die "Can not open $source for reading: $!";
open( OUT,    ">$out"    )
    or die "Can not open $out for writing: $!";

## parse source and blast chains
while ( <SOURCE> )
{
    chomp;
    next if ( /^\s*$/ );

    ## match lines
    if ( /^$DELIMITER/ )
    {

        ## a holder
        my @been_blasted;

        ## a little sanity
        next unless $id;

        ## up counter
        $pdb_count++;

        ## part of final counter
        print OUT "${ID}${id}\n";

        ## blast everything
        foreach my $i ( @chain_seq_pairs )
        {
            foreach my $j ( @been_blasted )
            {
                print OUT "${id} $i->[0] $j->[0]", '-' x 30, "\n";

                ## fill our blast i and j temp files
                open( BLASTI, ">$blast_i" )
                    or die "Can not open $blast_i for writing: $!";
                open( BLASTJ, ">$blast_j" )
                    or die "Can not open $blast_j for writing: $!";
                print BLASTI $i->[1];
                print BLASTJ $j->[1];
                close( BLASTI );
                close( BLASTJ );

                ## blast away
                my @blast_results = split( "\n", `$blast_exe $options 
+-i "$blast_i" -j "$blast_j"` );

                ## increase counter
                $blast_count++;

                ## now cycle through lines,
                ## put filters, what have yous here
                foreach my $line ( @blast_results )
                {
                    print OUT "${id} $i->[0] $j->[0]| $line\n";
                }
            }

            ## save current sequence to be blasted against
            push( @been_blasted, $i );
        }

        print OUT "\n";

        ## increase counters
        $chain_count    += scalar( @chain_seq_pairs );
        $wild_count    += scalar( @chain_wild );
        $stub_count    += scalar( @chain_stubby );
        $dna_count    += scalar( @chain_dna );

        ## clean up a few things
        undef @chain_seq_pairs;
        undef @chain_dna;
        undef @chain_stubby;
        undef @chain_wild;
        undef $id;
        undef $chain;

        next;
    }

    if ( /^$ID/ )
    {
        $id        = $';
        next;
    }

    ## These are noops, but this is how you can catch their values
    ## if ( /^$COMPOUND/ ) {}
    ## if ( /^$HEADER/ ) {}
    ## if ( /^$SOURCE/ ) {}

    if ( /^$CHAIN/ )
    {
        $chain    = $';
        next;
    }

    if ( /^$SEQUENCE/ )
    {

        my $tmp        = $';
        my $bad        = 0;

        if ( $tmp =~ /^x+$/i )
        {
            ## we're a straight wild card chain
            ## ignore because that means we can be anything
            ## just a bunch of aa's stuck together but
            ## we don't know / care about specifics
            push( @chain_wild, [ $chain, $tmp ] );
            $bad++;

        }

        if ( $tmp ne uc( $tmp ) )
        {
            ## we're dna, crude test, but it works on the data
            ## DNA is lower case
            push( @chain_dna, [ $chain, $tmp ] );
            $bad++;
        }

        if ( length( $tmp ) <= $min_size )
        {
            ## we're too short
            push( @chain_stubby, [ $chain, $tmp ] );
            $bad++;
        }

        next if $bad;

        push( @chain_seq_pairs, [ $chain, $tmp ] );

        next;
    }

}

## clean up
close(   OUT  );
close( SOURCE );
unlink( $blast_i, $blast_j );    ## no error checking?!

my $end            = time - $begin;

print <<"ENDOFPRINT";
PDBs found            : $pdb_count
Chains found            : $chain_count
Blasts performed        : $blast_count

    Wild sequences        : $wild_count
    Stubby sequences    : $stub_count
    DNA sequences        : $dna_count

TIME    : $end seconds
ENDOFPRINT

print "\nDone\n";
exit(0);

Again, this script should help you get started in parsing a PDB file. But, you will have to read up on DBI if you want to stuff the data in Access( there is plenty of material already on the site about DBI, btw ).

If you need some more help, feel free to send me an email.

Will perl for money
JJ Knitis
[phone]
[e-mail]

[esswired]

I cannot find a reference to "camel book" Can you help? Thanks

[pfaut]

The "camel book" is O'Reilly's "Programming Perl" ISBN 0-596-00027-8. No perl programmer should be without one.

---
print map { my ($m)=1<<hex($_)&11?' ':'';
$m.=substr('AHJPacehklnorstu',hex($_),1) }
split //,'2fde0abe76c36c914586c';

[Jaap]

Welcome to PerlMonks, esswired.
Have you succeeded in getting ANY perl script to work?
Have you tried running through the file using perl?

[esswired]

OK. I guess I am in over my head. I have a Palm PDB file & want to read the whole thing & want to change it into a CSV file or something readable for Excel. The only script which I have run successfully is the example (a .pl file) one sent with perl. It does not require any input. I would want to pass the name of the file I want to read. Does the code you sent have to be compiled? The PALM:PDB file I copied from CPAN seems to be a .pm file. Is the Camel book something I have? I found lots of references, but no book. Should I give up now?