#!/usr/bin/perl -w
#
# Claudia Gahleitner 0556004
#
# Exercise 5

use strict;
use warnings;

if ($#ARGV != 1)
{
        die ("USAGE: ex5.pl input.txt output.txt\n");
}

my $idx = 0;
my $input = $ARGV[$idx++];
my $output = $ARGV[$idx++];

open(INFO,   "<  $input")    || die("kann coding.fa nicht öffnen: $!");
open(OUTFILE, "> $output") or die "Kann output.txt nicht öffnen: $!";

my @coding = <INFO>;
my $i;
my $j;
my @newcoding;

chomp @coding;
foreach $i (@coding)
{
    if ($i =~ /^[A|G|C|T]+$/i)
    {
        push (@newcoding,$i);
    }
    elsif ($i =~ /^>(.+)/ && @newcoding)
    {
        my $mrna = transcript(@newcoding);
        my @protein = translate($mrna);
        print "$1: @protein\n";
        @newcoding = [];
    }
}


my $mrna = transcript(@newcoding);
#print "mrna: $mrna\n";

my @protein = translate($mrna);

sub transcript
{
    my @temp;
 foreach $i (@_)
 {
  $i =~ s/T/U/g;
  push (@temp,$i);
 }
 return reverse(join("", @temp));
}


sub translate
{
 my $mrna = pop(@_);

# print "now: $mrna\n";

 my %codonTable =  # associates base-triplets to the amino-acids they encode
 (
    "CUU" => "Leu", "CUC" => "Leu", "CUA" => "Leu", "CUG" => "Leu", "UUA" => "Leu", "UUG" => "Leu",
        "UCU" => "Ser", "UCC" => "Ser", "UCA" => "Ser", "UCG" => "Ser", "AGU" => "Ser", "AGC" => "Ser",
        "CGU" => "Arg", "CGC" => "Arg", "CGA" => "Arg", "CGG" => "Arg", "AGA" => "Arg", "AGG" => "Arg",
        "GGU" => "Gly", "GGC" => "Gly", "GGA" => "Gly", "GGG" => "Gly",
        "GCU" => "Ala", "GCC" => "Ala", "GCA" => "Ala", "GCG" => "Ala",
        "GUU" => "Val", "GUC" => "Val", "GUA" => "Val", "GUG" => "Val",
        "ACU" => "Thr", "ACC" => "Thr", "ACA" => "Thr", "ACG" => "Thr",
        "CCU" => "Pro", "CCC" => "Pro", "CCA" => "Pro", "CCG" => "Pro",
        "AUU" => "Ile", "AUC" => "Ile", "AUA" => "Ile",
        "UAU" => "Tyr", "UAC" => "Tyr",
        "UUU" => "Phe", "UUC" => "Phe",
        "UGU" => "Cys", "UGC" => "Cys",
        "AAU" => "Asn", "AAC" => "Asn",
        "GAU" => "Asp", "GAC" => "Asp",
        "CAA" => "Gln", "CAG" => "Gln",
        "GAA" => "Glu", "GAG" => "Glu",
        "CAU" => "His", "CAC" => "His",
        "AAA" => "Lys", "AAG" => "Lys",
        "AUG" => "Met",
        "UGG" => "Trp",
        "UAA" => "___", "UAG" => "___", "UGA" => "___",
        );

 #  every protein starts with Methinin (AUG)
 my $start = index($mrna, "AUG");
 if ($start == -1)
 {
     print "$mrna\n";
  die("ERROR: no start-codon found. Bailing out!");
 }

 my @protein;
 while (length($mrna) > 3)
 {
  my $codon = substr($mrna, $start, 3);
  $mrna = substr($mrna, 3);

  if (!exists($codonTable{$codon}))
  {
   die ("ERROR: $codon does not encode a valid amino acid. Bailing out!");
  }

  my $amino = $codonTable{$codon};
  if (length($codon) < 3 || !$amino)
   {
   die("ERROR: no stop-codon found.");
   }
  elsif ($amino eq "___") # we hit a stop-codon
   {
   last;
   }
  {
  push(@protein, $amino);
  }
 }
 return @protein;
}

close(INFO);
close(OUTFILE);