delangel 5a7444e186 First step in refactoring UG way of storing indel likelihoods - main motive is that rank sum annotations require per-read quality or likelihood information, and even the question "what allele of a variant is present in a read" which is trivial for SNPs may not be that straightforward for indels. ... This step just changes storage of likelihoods so now we have, instead of an internal matrix, a class member which stores, as a hash table, a mapping from pileup element to an (allele, likelihood) pair. There's no functional change aside from internal data storage. As a bonus, we get for free a 2-3x improvement in speed in calling because redundant likelihood computations are removed. Next step will hook this up to, and redefine annotation engine interaction with UG for indel case. git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5809 348d0f76-0448-11de-a6fe-93d51630548a		2011-05-15 23:04:11 +00:00
..
config	Provide a default logger, some config settings, and some doc updates.	2009-04-29 02:06:05 +00:00
src	First step in refactoring UG way of storing indel likelihoods - main motive is that rank sum annotations require per-read quality or likelihood information, and even the question "what allele of a variant is present in a read" which is trivial for SNPs may not be that straightforward for indels.	2011-05-15 23:04:11 +00:00
test	Cleanup GATKBAMIndex unit test to allow a more efficient access pattern for	2011-05-12 21:50:33 +00:00