Hmm, not good. Fixing the math in PBT resulted in changed MD5s for integration tests that look like significant changes. I am reverting and will report this to Laurent.

This commit is contained in:
Eric Banks 2012-08-16 21:41:18 -04:00
parent 65c594afff
commit 53383e82ec
1 changed files with 1 additions and 1 deletions

View File

@ -645,7 +645,7 @@ public class PhaseByTransmission extends RodWalker<HashMap<Byte,Integer>, HashMa
bestChildGenotype.clear();
bestChildGenotype.add(childGenotype.getKey());
}
else if(MathUtils.compareDoubles(configurationLikelihood, bestConfigurationLikelihood) == 0) {
else if(configurationLikelihood == bestConfigurationLikelihood) {
bestFirstParentGenotype.add(firstParentGenotype.getKey());
bestSecondParentGenotype.add(secondParentGenotype.getKey());
bestChildGenotype.add(childGenotype.getKey());