Functional equivalent of GCG Findpatterns?

Alan Bleasby ableasby at hgmp.mrc.ac.uk
Tue Jan 18 14:19:09 EST 2000


Regarding the FUZZies: They accept PROSITE style pattern specifications.
Ergo:

   A-[IL]-{P}-x(10)-v(2,5)-C

is OK. It is case insensitive and hypens are optional. You can specify
'<' for N terminus and '>' for C terminus

Fuzznuc is the same except N is used for unknows and ambiguity codes
or [] can be used but ambiguity codes are not allowed in the []'s.

Both of these programs analyse the pattern and chose one of seven or
so different algorithms which will be the speediest of those coded
for the given pattern. So there's Boyer-Moore-Horspool, Shift OR,
modified SO, Tarhio-Ukonnen, Baeza-Yates-Perleberg etc in there.


HTH

Alan





More information about the Bio-soft mailing list