How do I remove a part of the header of a fasta file in perl?

[Katty1]

I did a script in Perl that breaks several sequences of a multifasta file, but I need remove a part of string of header.

For example:

input file:
>gi|983431797|ref|NZ_LN868938.1| Nocardia farcinica genome assembly NCTC11134, chromosome : 1
CTGACTGGGAGTACGAAGGCCGCCTGCACAAGACAACGGGGCAGCGAACCTTCTTCTGCACCGGCACGGA
CGACGCCGAGATGCCTCGACCTGGAGAACCTCGGCCGCGGCGAACCGCTCGCCCATGTCCGCGCCGAGTT

Output file:
>Nocardia farcinica genome assembly NCTC11134, chromosome : 1
CTGACTGGGAGTACGAAGGCCGCCTGCACAAGACAACGGGGCAGCGAACCTTCTTCTGCACCGGCACGGA
CGACGCCGAGATGCCTCGACCTGGAGAACCTCGGCCGCGGCGAACCGCTCGCCCATGTCCGCGCCGAGTT

I've done everything, but I can't think of a solution to this.
Can someone help me?

My script:

#!/usr/bin/perl

use strict;
use warnings;
use IO::File;

my $file = "\nFILE: perl $0 <Fasta>"."\n";
print $file and exit unless($ARGV[0]);
my $input = IO::File->new("$ARGV[0]"), my $output;
while(my $line = $input->getline){
chomp($line);
if($line =~ /^>/){
$line =~ s/^>//;
$output = IO::File->new("> $line.fa");
print $output ">".$line."\n";
}else{
print $output $line."\n";
}
}
close($input);
close($output);

[SNPsaurus]

Can you just split on " " and print the second element of the split array? Or split on '|' and take the last element. It just depends on how standard the formatting is of the header.
@header_split = split(" ",$line);
$changed_line = $header_split[1];

[kmcarr]

Originally Posted by Katty1

I did a script in Perl that breaks several sequences of a multifasta file, but I need remove a part of string of header.

For example:

input file:

>gi|983431797|ref|NZ_LN868938.1| Nocardia farcinica genome assembly NCTC11134, chromosome : 1
CTGACTGGGAGTACGAAGGCCGCCTGCACAAGACAACGGGGCAGCGAACCTTCTTCTGCACCGGCACGGA
CGACGCCGAGATGCCTCGACCTGGAGAACCTCGGCCGCGGCGAACCGCTCGCCCATGTCCGCGCCGAGTT

Output file:

>Nocardia farcinica genome assembly NCTC11134, chromosome : 1
CTGACTGGGAGTACGAAGGCCGCCTGCACAAGACAACGGGGCAGCGAACCTTCTTCTGCACCGGCACGGA
CGACGCCGAGATGCCTCGACCTGGAGAACCTCGGCCGCGGCGAACCGCTCGCCCATGTCCGCGCCGAGTT

Katty,

I would caution you that what you plan to do is potentially problematic. The generally accepted format for FASTA file deflines is that the first word after the ">" represents the unique identifier for the sequence. The "first word" is defined as everything up to the first "whitespace" which may be a space or tab character. Everything that comes after that is optional description text. If you also have included in your analysis:

>gi|873551602|emb|LN868939.1| Nocardia farcinica genome assembly NCTC11134, plasmid : 2
GGCTTTGTGCCCGCCGAAAAAAGGTTGCCTATGTCCAAGCCTGCATTTACCGAAATCGACCGAATGACGG
GCGGAGGGCGGAGTAATCGCACCCGCCCACCGGTCAACTTCCTTCTTCACACCGAGGAAGGAAACTCGAG...

Which you also edit to:

>Nocardia farcinica genome assembly NCTC11134, plasmid : 2
GGCTTTGTGCCCGCCGAAAAAAGGTTGCCTATGTCCAAGCCTGCATTTACCGAAATCGACCGAATGACGG
GCGGAGGGCGGAGTAATCGCACCCGCCCACCGGTCAACTTCCTTCTTCACACCGAGGAAGGAAACTCGAG...

You have two sequences as part of your analysis which share the same ID, "Nocardia".

[GenoMax]

One alternative would be to change all spaces to "_" so that you have a long string (that should stay unique) for each fasta header. It would be cumbersome but would at least avoid the problem @kmcarr pointed out.