Hello reebee3,

Here’s one way to approach this task:

#! perl
use strict;
use warnings;

my (%seqs, $id, $dna);

while (my $line = <>)
{
    chomp $line;

    if ($line =~ / ^ > (.+) /x)
    {
        $seqs{$id} = $dna if defined $id;
        $id        = $1;
        $dna       = '';
    }
    else
    {
        $dna      .= $line;
    }
}

$seqs{$id} = $dna if defined $id;

for my $key (sort { length $seqs{$a} <=>
                    length $seqs{$b} } keys %seqs)
{
    printf "%s:%d\n", $key, length $seqs{$key};
}
[download]

Output:

15:55 >perl 1406_SoPW.pl data.fas
SequenceID|9876_Gene2:15
SequenceID|1234_Gene1:16

15:55 >
[download]

Notes:

• The above code contains no error checking! In particular, it doesn’t check that the fasta file format is valid. You say “I do not want to use BioPerl”, but a dedicated module is usually better and safer than hand-written code.
• The special filehandle <> reads from the file(s) specified on the command line (or from standard input if no files are specified). For other approaches, see perlopentut#Opening-Text-Files-for-Reading.
• You say you want to sort the data by length, but you don’t specify the sort order. I have assumed increasing order. If you want decreasing order instead, reverse the occurrences of $a and $b: sort { length $seqs{$b} <=> length $seqs{$a} }

Hope that helps,

G'day reebee3,

To read lines from a file on the command line, you can just do this:

while (<>) {
    # $_ holds the line read - process it here
}
[download]

In your foreach loop, you can get each key in the sorted order you want by changing

... (keys ...
[download]

to

... (sort {length $sequences{$a} <=> length $sequences{$b} } keys ...
[download]

With this input file:

$ cat pm_1144803_fasta.txt
>SequenceID|1234_Gene1
CTTTTAAGCTGATTAGGCTTTTATACCATTAGATTTAGTAACTATTGTCTTTTAA
>SequenceID|9876_Gene2
GTGCTGTCTTAAGTTGAACAGAGTGTGGGAGGAAATATAAGCAAAGTTATTCCGTAGAATT
>SequenceID|4567_Gene3
CATCCTCCTTTACACCCCACAAACATTTGGCAACCCCTGATAGGTTTCTTTCTTGTGGA
[download]

and this script:

$ cat pm_1144803_fasta_seq_len_sort.pl
#!/usr/bin/env perl

use strict;
use warnings;

my (%sequences, $seq_key);

while (<>) {
    chomp;

    if (/^>/) {
        $seq_key = substr $_, 1;
    }
    else {
        $sequences{$seq_key} = $_;
    }
}

foreach my $key (
    sort {length $sequences{$a} <=> length $sequences{$b} } keys %sequ
+ences)
{
    my $len = length ($sequences{$key});
    print "$key:$len\n";
}
[download]

You can run this from the command line:

$ pm_1144803_fasta_seq_len_sort.pl pm_1144803_fasta.txt
SequenceID|1234_Gene1:55
SequenceID|4567_Gene3:59
SequenceID|9876_Gene2:61
[download]

Very decent answers above. So this is merely a quibble.

You say

"I do not want to use BioPerl."

Why? ... and why should we not treat such an observation as being of the same order as
"I do not want to use a module." or even,
"I want to dig a hole but I do not want to use a shovel." ?

Using a shovel is usually easier than digging with your bare hands. Using a module is usually easier than reinventing something with its capabilities ... and using BioPerl (seems to be) a well-tested and effective method of achieving your goal.