[Bioperl-l] Checking molecular type for blast searches

simon andrews (BI) simon.andrews@bbsrc.ac.uk
Thu, 18 Apr 2002 13:38:24 +0100


> -----Original Message-----
> From: Jason Stajich [mailto:jason@cgt.mc.duke.edu]
> Sent: 18 April 2002 13:02
> Subject: Re: [Bioperl-l] Checking molecular type for blast searches
> 
> 
> In bioperl we guess it is DNA if the % of ATGC is > 85%. See
> Bio::PrimarySeq method _guess_alphabet for an impementation of this.

Just a little heads-up from our experience, we have found that checking the
% ATGCN is a better indication of sequence type.  A lot of DNA sequences
people are using these days contain poly-N sections (eg Ensembl chromosome
segments, or seqs which have been masked for repeats).  This is especially
true for sequences being passed to a blast search.

	TTFN

	Simon.

--
Simon Andrews PhD
Bioinformatics Dept
The Babraham Institute

simon.andrews@bbsrc.ac.uk
+44 (0)1223 496463