#!/usr/bin/perl -w
use strict;
#Code to calculate the % composition of aa in protein sequences in a multiple fasta file
#and save in .arff file format for input into Weka software
if (@ARGV !=3){
die "Supply a multi fasta file, output file (.arff) and class attribute (i.e.1 or 0)\n";
}
my $in_name=$ARGV[0];
my $out_name=$ARGV[1];
my $class = $ARGV[2];
my $sequence='';
my $aa='';
my @aa_code = qw /A C D E F G H I K L M N P Q R S T V W X Y/; 
my @cnt;
my $len=0;

#Opening files
open(FH1,"$in_name");
open(FH2,">$out_name");

printf FH2 "\@RELATION aa_comp\n";
for ($a=0;$a<21;$a++){printf FH2 "\@ATTRIBUTE aa_%s NUMERIC\n", $aa_code[$a];} 
printf FH2 "\@ATTRIBUTE class {1,0}\n";
printf FH2 "\@data\n";

$/ = "\n>"; #input_record separator
while($sequence=<FH1>)
{
  my @seq = ' ';
  chomp($sequence);
  $sequence =~ s/^>*.+\n//; #remove FASTA header
  $sequence =~ s/\n//g; #remove end lines
  $len=length($sequence);
  #exploding string into an array
  @seq = split('',$sequence);
  #setting count array values to zero for each sequence
  for ($a=0;$a<21;$a++){$cnt[$a]=0;}
  #counting up numbers of each type of aa in sequence
  for($a=0;$a<=$#seq;$a++){
    my $i=0;
    foreach $aa (@aa_code){
      #printf "matching aa $aa -- i is $i\n";
      if($seq[$a] eq $aa) {
	$cnt[$i]++;
	last;
      }
      $i++; 
    }
  }	
  #calculating aa compositions and printing out 
  for($a=0;$a<21;$a++){
    printf FH2 "%4.2f,", ($cnt[$a]/$len)*100;
  }
printf FH2 "%d \n", $class;
}
close FH1;
close FH2;