In article <49t1e8$mco at mark.ucdavis.edu>, ez041797 at dale.ucdavis.edu
(Jean-Manuel Henry) wrote:
> Hi, I am trying to determine heterologous nucleotides in a set of
> related sequences. So far, I have been able to determine homologous
> nucleotides, by performing a multiple sequence alignment and shading
> homologouse regions,2 this isn't of much use since the remaining
> sequences may hav homology with one another. For example:
> sequence 1: A
> sequence 2: A
> sequence 3: A
> sequence 4: C
> sequence 5: C
> sequence 6: T
> Conventional shading programs would shade all sequences containing "A"
> (1,2, & 3). However, I would like a program which would shade "C" as well
> in sequences 4 & 5. That is, a program to detect all nonunique
> nucleotides (T in sequence 6, in the above example). If anyone knows of a
> program with this ability please email me. Much obliged,
> JMH
I know of no progarm that will do this. However, you might want to check
out the BlockMaker server on the WWW at http://www.blocks.fhcrc.org. If
you provide this program with a fasta listing of your related proteins, it
will seek out regions of high homology (blocks). You can then get a
graphical representation of these regions in logo format (provided as a
postscript file). This sequence logo will give you a way to visualize all
bases that occur at a given position and there relative frequency. It's a
little tough to describe, give it a look...
Hope this helps.
Mike Coyne
mjcoyne at warren.med.harvard.edu
