Parker Method for Prediction of surface sites

Parker et al. has produced following hydrophilicity scale derived from High-Performance Liquid Chromatography (HPLC) Peptide Retention Data.

Surface Profile: The surface profile for a protein was determined by summing the parameters for each residue of seven-residue segment and assigning this sum to the fourth residue. This procedure was repeated by shifting the segment by one residue from the N-to the C-terminus. A plot of these values against the residue number using either HPLC, accessiblity, bulk hydrophobic character, hydrophobicity, global, pi, or hydropathy parameters (Here only HPLC is used) provided a surface profile.

To objectively interpret all of these profiles, Parker et al. has used following arbitary of rules:

1)  The average surface hydrophilicity is defined as the the mean of the profile values for a protein using a particular parameter set.
2) Any residue with a profile value greater than 25% above the average surface hydrophilicity value were defined as surface sites.

Program:

Parker.pl

use strict;
use warnings;
use FindBin qw($Bin);
use lib "$Bin";
use ParkerModule;

my %protein = ("A"=>{"Code"=>"Ala","Params"=>2.1},
               "R"=>{"Code"=>"Arg","Params"=>4.2},
               "N"=>{"Code"=>"Asn","Params"=>7.0},
               "D"=>{"Code"=>"Asp","Params"=>10.0},
               "C"=>{"Code"=>"Cys","Params"=>1.4},
               "E"=>{"Code"=>"Glu","Params"=>7.8},
               "Q"=>{"Code"=>"Gln","Params"=>6.0},
               "G"=>{"Code"=>"Gly","Params"=>5.7},
               "H"=>{"Code"=>"His","Params"=>2.1},
               "I"=>{"Code"=>"Ile","Params"=>-8.0},
               "L"=>{"Code"=>"Leu","Params"=>-9.2},
               "K"=>{"Code"=>"Lys","Params"=>5.7},
               "M"=>{"Code"=>"Met","Params"=>-4.2},
               "F"=>{"Code"=>"Phe","Params"=>-9.2},
               "P"=>{"Code"=>"Pro","Params"=>2.1},
               "S"=>{"Code"=>"Ser","Params"=>6.5},
               "T"=>{"Code"=>"Thr","Params"=>5.2},
               "W"=>{"Code"=>"Trp","Params"=>-10.0},
               "Y"=>{"Code"=>"Tyr","Params"=>-1.9},
               "V"=>{"Code"=>"Val","Params"=>-3.7},
              );
my $sequence = "MAFSAEDVLK EYDRRRRMEA LLLSLYYPND RKLLDYKEWS PPRVQVECPK
APVEWNNPPS KGLIVGHFS GIKYKGEKAQ ASEVDVNKMC CWVSKFKDAM
RRYQGIQTCK IPGKVLSDLD AKIKAYNLTV EGVEGFVRYS RVTKQHVAAF
LKELRHSKQY ENVNLIHYIL TDKRVDIQHL EKDLVKDFKA LVESAHRMRQ 
GHMINVKYIL YQLLKKHGHG PDGPDILTVK TGSKGVLYDD SFRKIYTDLG
WKFTPL";
sub clean_input_sequence {
 my $seq = shift;
 $seq =~s/\n//g;
        $seq =~s/ +//g;
 return $seq;
}

my $seq = clean_input_sequence($sequence);
my $windowlength = 7;

my $parker = ParkerModule->new(\%protein, $seq, $windowlength);
$parker->display_object();
my %seqplot = $parker->plotgraph;
sub asc_sort_subject {
 $a<=>$b;
}
my $key;
foreach $key(sort asc_sort_subject(keys %seqplot)){
   print "$key\t$seqplot{$key}{\"code\"}\t$seqplot{$key}
                {\"assignedvalue\"}\n"; 
}

ParkerModule.pm

package ParkerModule;
use strict;
use warnings;
use Math::BigFloat;

sub new {
 my $class = shift;
 my ($parameters,$sequence,$windowlength) =@_;
 my $ref={
  "Parameters"=>$parameters,
  "Sequence"=>$sequence,
  "Windowlength"=>$windowlength,
 };
 bless($ref,$class);
 return $ref;
}

sub display_object {
 my $self = shift;
 #print "Parameters= ".$$self{"Parameters"};
 my $param = $$self{"Parameters"};
 #print $$param{"A"}{"Params"},"\n";
 #print "Sequence= ".$$self{"Sequence"};
 #print "Windowlength= ".$$self{"Windowlength"};
  
}

sub plotgraph {
 my $self = shift;
 my $sequence = $$self{"Sequence"};
 my @seq = split("",$sequence);
 my $param = $$self{"Parameters"};
 my %seqplot;
 
 my $length = @seq;
 my $windowlength = $$self{"Windowlength"};
 my $meanposition = int($windowlength/2);
 for(my $i=0;$i<= $length-$windowlength; $i++ ){
  my $sum = 0;
  my $s = Math::BigFloat->new($sum);
  for(my $j = $i; $j< $i+ $windowlength; $j++){
   $s = $s + $$param{"$seq[$j]"}{"Params"};
   
  }
  
  my $k = $i+$meanposition;
  $seqplot{"$k"}{"assignedvalue"} = $s;
  $seqplot{"$k"}{"code"}= $seq[$k];
 }
 
 return %seqplot;
}

1;

References:
New Hydrophilicity Scale Derived from High-Performance Liquid Chromatograpy Peptide Retention Data: Correlation of Predicted Surface Residues with Antigenicity and X-ray-Derived Accessible Sites
J.M.R.Parker, D.Guo, and R.S.Hodges

Please go through the program and reference, and suggest error and improvement of the program.