Re: How to get non-redundant DNA sequences from a FASTA file?

[Athanasius]

Hello supriyoch_2008,

Same approach as choroba, except that I read the data into a hash mapping headers to sequences, and then reverse that hash to get the non-redundant sequences:

#! perl
use strict;
use warnings;

my $fasta =
'>gi1 cds
ATG fun 
>gi2 cds
ATG fun
>gi3 cds
GGG fun';

my %hdrs;
   $hdrs{$1} = $2 while $fasta =~ / > (.+) \s+ cds \s+ (.*) \s+ fun /g
+x;

print  " A. Header & sequences are:\n";
printf ">%s cds\n%s\n", $_, $hdrs{$_} for sort keys %hdrs;

my %seqs;

while (my ($key, $value) = each %hdrs)
{
    push @{$seqs{$value}}, $key;
}

print  " B. Only sequences are:\n";
printf "$_\n" for sort keys %seqs;

print " C. Non-redundant sequences are:\n";
printf ">%s cds\n%s\n", ( sort @{$seqs{$_}} )[0], $_ for sort keys %se
+qs;

Output:

18:47 >perl 1009_SoPW.pl
 A. Header & sequences are:
>gi1 cds
ATG
>gi2 cds
ATG
>gi3 cds
GGG
 B. Only sequences are:
ATG
GGG
 C. Non-redundant sequences are:
>gi1 cds
ATG
>gi3 cds
GGG

18:47 >

Note: on hash reversal, see How-do-I-look-up-a-hash-element-by-value of perlfaq4.

If you really don’t care which header is output when there are redundant sequences, you can just say:

my %seqs = reverse %hdrs;
...
print " C. Non-redundant sequences are:\n";
printf ">%s cds\n%s\n", $seqs{$_}, $_ for sort keys %seqs;

When your code starts to get too complicated, it’s usually a good idea to step back and look for a simpler approach. Remember, less is more. ;-)

Hope that helps,

[supriyoch_2008]

Hi Athanasius,

Thank you for your suggestions and fixing the problem. Your script has worked nicely. I am grateful to you.

With regards,