On Fri, Jan 24, 2014 at 2:29 AM, Gregory Ewing <greg.ew...@canterbury.ac.nz>wrote:
> If all you want is a mapping between a sequence of > length n and compact representation of it, there's > a much simpler way: just convert it to a base-k > integer, where k is the size of the alphabet. > > The resulting integer won't be any larger than an > index into the de Bruijn sequence would be, and > you can easily recover the original sequence from > its encoding without needing any kind of lookup > table. > True, the "all you want is a mapping" is not quite true. I actually plan to plot frequency (the number of times an observed sub sequence overlaps a value in the De Bruijn sequence) The way the sub sequences overlap is important to me and I don't see a way go from base-k (or any other base) to the index location in the De Bruijn sequence. i.e. a decoding algorithm. Vincent Davis 720-301-3003
-- https://mail.python.org/mailman/listinfo/python-list