gatk-3.8/java/test/data/tranches.4.txt

FDRtranche,novelTITV,pCut,numNovel,filterName
0.10,2.8000,1.1800,893,FDRtranche0.00to0.10
1.00,2.7782,1.0600,903,FDRtranche0.10to1.00
10.00,2.5714,0.3800,975,FDRtranche1.00to10.00
100.00,1.6091,0.0000,1602,FDRtranche10.00to100.00
Initial commit, without any tool changes, of a new infrastructure for determining tranches. This new version walker up from the lowest quality snps and determines Ti/Tv. This is marginally more stable than moving in the other direction when there are few novel variants (exomes). Can make a substantial difference in the size of the call set (10-20%). I'll hook it into the main system now. Includes an new class Tranche, isolated read/writing utilities that are now testing in TestVariantRecalibrator, which should be moved to UnitTest as soon as I can figure out how to do this on my mac. git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@4654 348d0f76-0448-11de-a6fe-93d51630548a 2010-11-12 07:52:49 +08:00			`FDRtranche,novelTITV,pCut,numNovel,filterName`
			`0.10,2.8000,1.1800,893,FDRtranche0.00to0.10`
			`1.00,2.7782,1.0600,903,FDRtranche0.10to1.00`
			`10.00,2.5714,0.3800,975,FDRtranche1.00to10.00`
			`100.00,1.6091,0.0000,1602,FDRtranche10.00to100.00`