WG: [Bioperl-l] dssp script

Robinson, Peter peter.robinson at charite.de
Mon Feb 14 05:30:16 EST 2005


Thanks Ed,

the iterator seems to work perfectly. The module is quite useful for me. It may be that I am doing something wrong, but it seems as if there may be one remaining bug. If you look at line 1206 $key_num = $arg_str;
For some files it seems as if $key_num gets assigned a string such as "110:" rather than the integer 110. This messes us line 1215 
if ($key_num == $self->{'Res'}->[$i]-{'pdb_resnum'})
I added 
$key_num =~ s/://;
after line 1206, which fixed the problem. 

Thanks again, peter
Peter N. Robinson, M.D.
Institute of Medical Genetics
Charité University Hospital
Augustenburger Platz 1
13353 Berlin
Germany
++49-30-450 569124
peter.robinson at charite.de
http://www.charite.de/ch/medgen/robinson
The best editor for MS Word documents: /dev/null



-----Ursprüngliche Nachricht-----
Von: Ed Green [mailto:ed at compbio.berkeley.edu]
Gesendet: Dienstag, 1. Februar 2005 00:08
An: ed at compbio.berkeley.edu
Cc: Robinson, Peter; bioperl-l at portal.open-bio.org
Betreff: Re: [Bioperl-l] dssp script


Peter-
Just checked in these fixes to Bio::Structure::SecStr::DSSP:Res.pm
You may find the new residues() interator method useful.

Regards,
Ed Green


Ed Green wrote:
> Dear Peter,
> 
> These two are in fact bugs that I will fix. The first results because of 
>  the presence of 'termination residues' that don't have residue numbers. 
> Their residue numbers, then, can't be compared numerically. Fortunately, 
> this bug won't result in wrong results as we want this comparison to 
> always be false anyway. The solution to this is to first check if either 
> of the termination residue signals are set and if so, don't do this 
> numerical comparison.
> 
> The second, blank line(s) at end of file will also be fixed.
> 
> Beware that there is, I think, a bug in your script. It appears that you 
> are attempting to iterate over all residues. However, iterating A:1 .. 
> A:max doesn't get it done because of the crazy way residues can be 
> numbered in PDB files: you'll miss all the residues with altloc codes 
> (A:27A, A:27B, A:27C, e.g.).
> 
> To make this easy an iterator is called for. It will just return all 
> 'real' residues for the pdb file or for a specified chain - I'll try to 
> get that done this weekend.
> 
> Regards,
> Ed Green
> 
> Robinson, Peter wrote:
> 
>> Dear BioPerlers,
>>
>> I am writing a script to use the BioPerl DSSP module to print out a 
>> list of phi and psi angles for all applicable residues  of all chains. 
>> Although the results are correct, I get the following error message at 
>> the end of each chain:
>>
>> Argument "" isn't numeric in numeric eq (==) at 
>> /usr/local/share/perl/5.8.4/Bio/Structure/SecStr/DSSP/Res.pm line 1168.
>>
>> and I am not quite sure where it is coming from. Perhaps I am using 
>> the wrong part of the API, but I am trying to get a list of all 
>> residues for each chain as follows:
>>
>> foreach my $ch (@chains) {
>>   my $ss_elements_pts = $dssp->secBounds($ch);
>>   print "Chain $ch:\n";
>>   my $pos = 0;
>>   my $max = 0;
>>   foreach my $stretch (@{$ss_elements_pts}) {
>>     my $start = $stretch->[0];
>>     my $end = $stretch->[1];     if ($end =~ m/(\d+)/) { $end = $1; }
>>        if ($end  > $max) { $max = $end; }
>>   }
>>   ## END is now the last residue in this chain
>>   for my $res (1..$max) {
>>     my $residueID = $res . ":" . $ch;
>>     my ($phi,$psi,$SS,$SSsum,$AA);
>>     eval { $phi = $dssp->resPhi($residueID);};
>>     etc.
>>
>> The full script is appended to the bottom of this mail.
>>
>>
>> I also noticed what might be a minor bug in the module DSSP/Res.pm; 
>> when I use dsspcmbi to analyze a PDB file, it produces a results file 
>> with an empty last line. This causes a crash:
>>
>> Use of uninitialized value in chomp at 
>> /usr/local/share/perl/5.8.4/Bio/Structure/SecStr/DSSP/Res.pm line 
>> 1284, <GEN1> line 955.
>>
>>
>>  If I manually remove this last empty line, there was no error. By 
>> adding the following line at Res.pm l.1284, you can fix the problem:
>>
>>
>>  while ( chomp( $cur = <$file> ) ) {
>>       next if ($cur =~ m/^\s*$/);  
>> *********************************************
>>     $res_num = substr( $cur, 0, 5 );
>>     $res_num =~ s/\s//g;
>>     $self->{ 'Res' }->[ $res_num ] = &_parseResLine( $cur );
>>     }
>> }
>>
>>
>>
>>
>> Thanks in adavance for any tips! Peter
>> Peter N. Robinson, M.D.
>> Institute of Medical Genetics
>> Charité University Hospital
>> Augustenburger Platz 1
>> 13353 Berlin
>> Germany
>> ++49-30-450 569124
>> peter.robinson at charite.de
>> http://www.charite.de/ch/medgen/robinson
>> Beware of bugs in the above code; I have only proved it correct, not 
>> tried it. -Donald Knuth, computer scientist (1938- )
>>
>> ########################
>>
>> #!/usr/bin/perl -w
>> use IO::File;
>> use Bio::Structure::SecStr::DSSP::Res;
>> use Data::Dumper;
>>
>>
>> =pod
>> parseDSSP.pl
>> Script to parse the output of DSSP using the BioPerl module
>> Bio::Structure::SecStr::DSSP::Res. To use it, process a PDB
>> file with dssp or dsspcmbi, and pass the resulting file to this 
>> script. For more information on dssp and BioPerl see the
>> module documentation at http://bioperl.org
>>
>> @email peter.robinson at charite.de
>> 21 January, 2005
>>
>> =cut
>>
>>
>>
>> my $file = "pdb43ca.dssp";
>> my $dssp = new Bio::Structure::SecStr::DSSP::Res('-file'=> "$file");
>>
>> my $pdbID = $dssp->pdbID();
>> my $auth  = $dssp->pdbAuthor();
>> my $cmpd = $dssp->pdbCompound();
>> my $pdb_date = $dssp->pdbDate();
>> my $header = $dssp->pdbHeader();
>> my $pdbSource = $dssp->pdbSource();
>>
>> print "PDB entry $pdbID \n\tauthor:\t$auth",
>>   "\n\tCompound:\t$cmpd",
>>   "\n\tDate:\t$pdb_date",
>>   "\n\tHeader:\t$header",
>>   "\n\tsource:\t$pdbSource\n\n";
>>
>> my $totalRes = $dssp->numResidues();
>> print "Total residue count (all chains):$totalRes\n";
>>
>>
>> my $surArea= $dssp->totSurfArea();
>> print "Total accessible surface area:\t$surArea  (square Ang)\n";
>>
>>
>> my $chainRef = $dssp->chains();
>> my @chains = sort  @{$chainRef};
>> print "Chain[s]:\n";
>> foreach my $ch (@chains) {
>>   print "\t$ch";
>> }
>> print "\n";
>>
>> my $hb = $dssp->hBonds();
>> print "H BONDS.\n";
>> print "TYPE O(I)-->H-N(J): $hb->[0]\n",
>>    "IN PARALLEL BRIDGES: $hb->[1]\n",
>>    "IN ANTIPARALLEL BRIDGES $hb->[2]\n",
>>    "TYPE O(I)-->H-N(I-5) $hb->[3]\n",
>>    "TYPE O(I)-->H-N(I-4) $hb->[4]\n",
>>    "TYPE O(I)-->H-N(I-3) $hb->[5]\n",
>>    "TYPE O(I)-->H-N(I-2) $hb->[6]\n",
>>    "TYPE O(I)-->H-N(I-1) $hb->[7]\n",
>>    "TYPE O(I)-->H-N(I+0) $hb->[8]\n",
>>    "TYPE O(I)-->H-N(I+1) $hb->[9]\n",
>>    "TYPE O(I)-->H-N(I+2) $hb->[10]\n",
>>    "TYPE O(I)-->H-N(I+3) $hb->[11]\n",
>>    "TYPE O(I)-->H-N(I+4) $hb->[12]\n",
>>    "TYPE O(I)-->H-N(I+5) $hb->[13]\n",
>>   "\n";
>>
>>     
>> foreach my $ch (@chains) {
>>   my $ss_elements_pts = $dssp->secBounds($ch);
>>   print "Chain $ch:\n";
>>   my $pos = 0;
>>   my $max = 0;
>>   foreach my $stretch (@{$ss_elements_pts}) {
>>     my $start = $stretch->[0];
>>     my $end = $stretch->[1];     if ($end =~ m/(\d+)/) { $end = $1; }
>>        if ($end  > $max) { $max = $end; }
>>   }
>>   ## END is now the last residue in this chain
>>   for my $res (1..$max) {
>>     my $residueID = $res . ":" . $ch;
>>     my ($phi,$psi,$SS,$SSsum,$AA);
>>     eval { $phi = $dssp->resPhi($residueID);};
>>     eval { $psi = $dssp->resPsi($residueID);};
>>     eval { $SS = $dssp->resSecStr($residueID);};
>>     eval { $SSsum = $dssp->resSecStrSum($residueID);};
>>     $AA = $dssp->resAA($residueID);
>>     $phi = $phi || "n/a";
>>     $psi = $psi || "n/a";
>>     $SS = $SS || "-";
>>     my $SSclass;
>>     if ($SSsum eq "H") { $SSclass = "helix"; }
>>     elsif ($SSsum eq "T") { $SSclass = "turn"; }
>>     elsif ($SSsum eq "B") { $SSclass = "beta"; }
>>     else { $SSclass = $SSsum; }
>>     print "$residueID) [$AA] phi:$phi psi:$psi SecStruct: $SS 
>> ($SSclass) \n";
>>   }
>> }
>>
>>
>>
>>
>>
>>
>>
>> _______________________________________________
>> Bioperl-l mailing list
>> Bioperl-l at portal.open-bio.org
>> http://portal.open-bio.org/mailman/listinfo/bioperl-l
> 
> _______________________________________________
> Bioperl-l mailing list
> Bioperl-l at portal.open-bio.org
> http://portal.open-bio.org/mailman/listinfo/bioperl-l



More information about the Bioperl-l mailing list