How To Write Sequences To Fasta Format With No Linebreaks (Wide Format) Using Bioperl

Question

[angel.roey]

I'm trying to export aligned sequences to a fasta file one by one using Bio::SeqIO. The result is that the sequences are broken by a new line every 60 columns. How do I avoid that?
I'd like to have the sequences exported in a 'wide' format, i.e. no line breaks in the sequence.

My code is roughly:

use Bio::SeqIO;
my $seqin = Bio::SeqIO->new(-file => "<$fastaFile", '-format' => 'Fasta');
my $outname = fileparse($fastaFile, qr/\.[^\.]*$/) . "_sub.fasta";
my $seqout = Bio::SeqIO->new(-file => ">$outname", '-format' => 'Fasta');

while(my $seq = $seqin->next_seq){
      # do something with $seq
      $seqout->write_seq($seq);
}

[Kenosis]

"...the sequences are broken by a new line every 60 columns. How do I avoid that?"

Why do you want to avoid that?

Answer 1

[kmcarr00]

Yes, there is a simple parameter for controlling the width of the output, called surprisingly -width.

my $seqout = Bio::SeqIO->new(-file=>"$outFile", -format=>'Fasta', -width=>"$lineWidthValue")

The value of -width appears to be limited to 32,766 (i.e. essentially 2^15) so it isn't possible to write infinitely long lines with this method.

If $lineWidthValue is not defined, or == 0 the output will default to 60 characters per line.

[kmcarr00]

Commenting on my own answer (bad form, yes) but it appears that it is possible using the -width method to get infinitely long lines. If $lineWidthValue == -1 the sequence will be output as a single line. There is a caveat however; if your pass '-1' to -width then write_seq does not print a newline at the end of the sequence as it normally would. You would need to manually print a newline to your output stream.

Answer 2

[fo3c]

I don't know bioperl, but I would circumvent the problem by opening $seqout as

open($seqout,'>',$outname);

and outputting as

print $seqout ">",$seq->header,"\n",$seq->sequence,"\n";

Do replace $seq->header and $seq->sequence with the appropriate variables/methods.

(don't forget to close $seqout)

[Kenosis]

Bio::SeqIO handles all file i/o; there's no need to explicitly open or close the fasta files. The notation open($seqout,'>',$outname); attempts to use a Bio::SeqIO object as a file handle. This shouldn't be done, since that object has its own methods.

[fo3c]

That's not what I meant. I suggested replacing the SeqIO way to output with a regular perl way to avoid the problem described in the question.

[angel.roey]

Yep, my way of circumventing it was exactly that -- open a file handle and print $seq->header and $seq->sequence to that file. But I thought there must be a more elegant way.
@Kenosis what would be the solution then? Is there a method for a wide fasta format in Bio::SeqIO?

[Kenosis]

emphasized textIs there a method for a wide fasta format in Bio::SeqIO?

I haven't found one--yet--either googling or by examining the Bio::SeqIO and Bio::Seq modules. The OP's also posted the question at StackOverflow with no responses as of the time of this reply.

[SES]

This is not true. Sometimes you want to pass a filehandle and not a filename, this is perfectly acceptable and there's a method for that, also :). For example, if you have the lexical filehandle $in, you'd pass the key/values of "-fh => \*$in" to the method new (instead of the familiar "-file => $filename").

[Kenosis]

Good correction! Appreciate it...