Re: Get random unique lines from file

[BrowserUk]

This script takes the name of the file and the number of sequences you want and outputs that number of randomly selected sequences.

It is simple, fast and should handle any size of input file with minimal memory usage:

#! perl -slw
use strict;
use List::Util qw[ shuffle ];

$/ = '>';

open FASTA, '<:raw', $ARGV[0] or die $!;
my @seqPosns;
push @seqPosns, tell( FASTA ) while <FASTA>;

@seqPosns = shuffle @seqPosns;

for ( @seqPosns[ 0 .. $ARGV[ 1 ] // 10 ] ) {
    seek FASTA, $_, 0;
    my $seq = <FASTA>;
    chomp( $seq ); chop( $seq );
    print '>', $seq;
}
close FASTA;

__END__

C:\test>988096.pl C:/dell/test/LCS/bioMan.fasta 2
>af418682
TTCCACAACTTTCCACCAAGCTCTACAAGATCCCAGAGTCAGGGGCCTGTATTTTCC
TGGGTCTTTTGGGCTTTGCCGCTCCATTTACACAATGTGGTTATCCTGCATTAATGC
ACTTCTTTCCTTCAGTACGAGATCTCCTAGATACCGCCTCAGCTCTATATCGGGAAG
TCAAACAATCCAGATTGGGACTTCAACCCCATCAAGGACCACTGGCCACAAGCCAAC
>ab033557
CTCCACGACATTCCACCAAGCTCTGCTAGATCCCAGAGTGAGGGGCCTTTACTTTCC
TGGGTCTTTTGGGCTTTGCTGCCCCTTTTACACAATGTGGCTATCCTGCCTTAATGC
ACTTCTTTCCTTCCATTCGAGATCTTCTCGACACCGCCTCTGCTCTGTATCGGGAGG
TCAAACAATCCAGATTGGGACTTCAACCCCAACAAGGATCAATGGCCAGAAGCAAAT
>x97850
CTCCACAACTTTCCTCCAAACTCTTCAAGATTCCAGAGTCAGGGCCCTGTACCTTCC
TGGGTCTTTTGGGGTTTGCCGCCCCTTTCACGCAATGTGGATATCCTGCTTTAATGC
ACTTTTTTCCTTCTATTCGAGATCTCCTCGACACCGCCTCTGCTCTGTATCGGGAGG
TCAGAAAATCCAGATTGGGACCTCAACCCGCACAAGGACAACTGGCCGGACGCCAAC

[Marshall]

I like your solution and ours are similar. You know more about the FASTA format than me and were able to "decode" the OP'S intent better than me. I like it!

I do admit confusion about this "// 10", I don't understand why that would be necessary? Or what the purpose is?


for ( @seqPosns[ 0 .. $ARGV[ 1 ] // 10 ] ) {
...blah....}

[BrowserUk]

It simply means that if I forget/omit the second parameter (the number of output records to produce), I get 10 by default.