[Biojava-l] Biojava-l Digest, Vol 99, Issue 12

[Andreas Prlic]

It sounds like you are talking about doing a needleman wunsch
alignment, while not penalizing end gaps. This should give you quite
similar results to just doing a
smith waterman one.

Is it the alignment display that you are concerned about? You could
easily fill up the ends with unaligned regions for smith waterman
results.  I don't find end-gaps particularly informative. Usually an
alignment display will give you the aligned positions, then you can
easily see (or script) if there are end gaps.

Hope that makes sense..
Andreas



On Fri, Apr 15, 2011 at 12:30 PM, Jay Vyas <jayunit100 at gmail.com> wrote:
> Consider Kalirin , a multidomain protein.
> http://www.ncbi.nlm.nih.gov/Structure/cdd/wrpsb.cgi?INPUT_TYPE=live&SEQUENCE=NP_003938.1
> Now, lets say I aligned Kalirin with a single domain member of the RHO-Gef
> family....  Then only one part of the sequence would match....  But what if
> I WANTED a global alignment, for programmatic purposes... Then I would want
> an alignment like this
> ------------------------------------------------RHOGEFRHOGEFRHOG--FRHOGEFRHOGEFRHOGEF------
> With lots of gaps in the beggining, and lots of matches at the end, where
> the length of the alignment between the two proteins was exactly equivalent.
>  Contrast this with a smith waterman output, which would look like this :
> RHOGEFRHOGEFRHOG--FRHOGEFRHOGEFRHOGEF....
> So is there a way to get the specificity of Smith Waterman in a Needlman
> Wunsch alignment ?