[Bioc-sig-seq] strandedness bug in Rsamtools?

[Steve Lianoglou]

On Sat, Feb 6, 2010 at 5:13 PM, Martin Morgan <mtmorgan at fhcrc.org> wrote:
> Thanks Steve -- there's a single line change to BSgenome (in the early
> lines of strand.R) that restores strand levels and behavior to
> Rsamtools, but we are still not fully aware of what consequences this
> might have for other packages so probably best to stick with your 'last
> known good' installation of early January. Certainly your flags should
> be parsed correctly.

For what it's worth, I tried simply changing the order of the factors
in that BSgenome::strand.R file and ran my tests -- I don't think that
solved the problem, though. Take that with a grain of salt, though ...
there's a good chance I can be mistaken. I'm debating even sending
this email because I'm not sure how thorough I tested that. I quickly
looked at some result I expected  and it seemed wrong ...

As I'm in "mad-dash" mode right now trying to get some things together
for a deadline I just took the "nuclear" option and backtracked to
earlier revisions of these packages.

Once I get out of the woods, I'll be happy to test more thoroughly. I
have some samples with sequences that all start with the same 4 NT's
on the 5' of the tag (a restriction enzyme was used to cut and create
"sequence tags", much like the MPSS protocols of "yore" :-)).

I put together some scratch code to test that scenarios (i) and (ii)
[from previous email] are kosher for these samples (ie. the correct 4
NT's are the 5' of the sequences in the entire experiment), so I can
"easily" check that as soon as I finish this work up.

-steve

-- 
Steve Lianoglou
Graduate Student: Computational Systems Biology
 | Memorial Sloan-Kettering Cancer Center
 | Weill Medical College of Cornell University
Contact Info: http://cbio.mskcc.org/~lianos/contact