[Bioperl-l] Bio::Seq::SequenceTrace and Bio::SeqIO::scf

[Chad Matsalla]

Tony Cox wrote:

> I just looked at the changes anf they don't confict with anything I
> have except my proposed change to the way the filehandles are used if
> it is a String::IO handle. I havn't committed these yet anyway - I'm
> trying to hunt down a 5.004 installation to test it on.

Sounds great Tony. I _completely_ rewrote scf.pm . The only things that
might cause grief for you are the templates used in the "pack"
statements. Those gave me grief to make work. Whew.

> Chad's stuff looks good to me. I'm curious as to why you might want to
> emulate a chromatogram in an scf file.

If you had, for example, a SequenceWithQuality and wanted to have a
false trace. Not that anything like consed would crash or refuse to
display a contig because a chromatogram was absent, or anything. Not at
all.

> The trace slice looks useful but I'm not sure whether one can use it
> to slice a "real" scf file, or only for home-made /emulated ones.

REAL! REAL! That was the point of me rewriting all of this stuff. The
code that I provided to the list yesterday uses a real scf from t/data
and the sub-scf is completely valid.

> Trimming real files would be v. useful for me since I have already
> written code to do this the long way while rendering multi-trace
> assemblies anchored about a single basepair.

This sounds excellent.

> I was seeing a problem with an off-by-one error in the SWQ module but
> I havn't tracked it down yet to either my code or the SWQ module.

I wrestled with an off-by-one for a while in phd.pm . I am unable to
reproduce it but I bet it is really in swq. Swine error.

Chad Matsalla


-- 
Dentist, n.:
	A Prestidigitator who, putting metal in one's mouth, pulls
	coins out of one's pockets.
		-- Ambrose Bierce, "The Devil's Dictionary"