[Bioperl-l] SeqIO::abi use

[Tom Keller]

Greetings,
Can the Bio::SeqIO::abi module be used to access the raw fluorescence data? Are there any examples available? I'm wondering how to calculate ratios of mixed bases prior to the smoothing and scaling done by the base caller.

thank you
Tom
kel...@ohsu.edu<mailto:kel...@ohsu.edu>
503-494-2442








_______________________________________________
Bioperl-l mailing list
Biop...@lists.open-bio.org
http://lists.open-bio.org/mailman/listinfo/bioperl-l


--
Subscription settings: http://groups.google.com/group/bioperl-l/subscribe?hl=en

[Roy Chaudhuri]

Hi Tom,

It looks from the docs/source like you can, with something like this:

use Bio::SeqIO;
my $seq=Bio::SeqIO->new(-file=>$ARGV[0], -format=>'abi',
-get_trace_data=>1)->next_seq;
my @atrace=$seq->get_trace_graph(-trace=>'a'); # or c,g,t

Sorry, but I don't have Staden installed so I can't test if this works.
You might also look at the non-BioPerl ABI module on CPAN - again, I
haven't tested this, but it looks like it could be useful for your purposes:
http://search.cpan.org/~MALAY/ABI/ABI.pm

Roy.

[Phillip San Miguel]

Hi Tom and Roy,
Not sure what the intention is here, but there are caveats you might
want to consider:
The problem with looking at the raw trace data is that it has not
yet been base called, no peaks have been found, nor have the dye-shifts
been compensated for, nor has any background been subtracted. So finding
the peaks in the raw data that correspond to the mixed base peak of
interest is not trivial. Then once you have, you would need to subtract
background (which can be different for each of the data channels.) Then
integrate the peaks.
Then what do you have? The % of the base at that position is A,C,G
or T?
I would not be so sure. The dyes may give different quantum yields,
the CCD may be more sensitive to some wavelengths than others, the
sequencing polymerase may prefer to add a certain base in your given
sequence context more than another. Really you would need to do standard
runs (mixing sequences in each of the base channels to be analyzed of
each base) in known ratios and create a standard curve to be sure.

Phillip

[Cook, Malcolm]

I don't think so, but c.f. http://search.cpan.org/~vita/Bio-Trace-ABIF-1.05/lib/Bio/Trace/ABIF.pm#raw_data_for_channel%28%29

Malcolm Cook
Stowers Institute for Medical Research - Bioinformatics
Kansas City, Missouri USA

[Joel Martin]

I'd use phred instead of kbcaller, and with the -d for .poly files
option, then parse the .poly files,
sounds like it might be the information you're after.

Joel