Hi,

I'd say you best stick with Ketil's second comment on having an
interface on an underlying sequence type. I'd go with that, too, as
tying-the-knot will not be the most efficient thing in terms of raw
speed.

But ultimately, it all depends on the target algorithm, circular folding
of RNA (for example) works on the raw input.

Gruss,
Christian

* Ketil Malde <[email protected]> [20.09.2011 09:19]:
> Dan Fornika <[email protected]> writes:
> 
> > Just off the top of my head, I'd think that using some form of the
> > repeat' function to (lazily) produce an endlessly-repeating sequence
> > could simulate a circular sequence.
> 
> Yes, that could work, although you'd still have to start somewhere (so
> if you back up far enough, you'll fall off the beginning).  You could
> also look at "tying the knot"-type of circular structures, but that
> would change the basic data structure.
> 
> The most direct (and perhaps pedestrian) way might be to just define an
> interface that, say, stores the sequence as a bytestring, but allows you
> to extract sequence fragments that span the origin for analysis.
> 
> -k
> -- 
> If I haven't seen further, it is by standing in the footprints of giants
> _______________________________________________
> Biohaskell mailing list
> [email protected]
> http://malde.org/cgi-bin/mailman/listinfo/biohaskell

Attachment: pgpp83ca9DPIn.pgp
Description: PGP signature

_______________________________________________
Biohaskell mailing list
[email protected]
http://malde.org/cgi-bin/mailman/listinfo/biohaskell

Reply via email to