On Oct 5, 2011, at 4:02 AM, Ketil Malde wrote: > Btw, would it be possible to make samtools not depend on seqloc? The > latter pulls in a load of dependencies, it's no big deal, but it seems > that samtools could be useful on its own? (Read: I'm using it without > using seqloc :-)
I could look in to it, but I do use the seqloc internally, especially for dealing with the Cigar string descriptor, and it's also used in the external interface. Looking at the seqloc dependencies, QuickCheck & random are both there just for testing and it would make sense to separate them. If I did that, the set of dependencies for seqloc would be pretty minimal--attoparsec (which pulls in containers and deepseq), biocore, and bytestring > Also, I have a small samtools-unsafe library, it just lets you extract > all alignments from the whole file or from a region as a lazy list of > Bam1s. This isn't very fashionable (*cough* iteratees *cough*), but I > find it rather convenient. Would you be interested in merging this into > samtools, or should I just keep it separate? That sounds like a useful interface to provide--I'll make an up-to-date version of the darcs repos available if you can then put together a patch to merge in your library. Best, --Nick _______________________________________________ Biohaskell mailing list [email protected] http://malde.org/cgi-bin/mailman/listinfo/biohaskell
