From ba2acda406d9c567ec1b9e954431a7e6cc4176fe Mon Sep 17 00:00:00 2001 From: rpoplin Date: Fri, 15 Jan 2010 18:36:14 +0000 Subject: [PATCH] Clarifying the comment regarding differentiating between first and second of pair in CycleCovariate. git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@2594 348d0f76-0448-11de-a6fe-93d51630548a --- .../sting/gatk/walkers/recalibration/CycleCovariate.java | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) diff --git a/java/src/org/broadinstitute/sting/gatk/walkers/recalibration/CycleCovariate.java b/java/src/org/broadinstitute/sting/gatk/walkers/recalibration/CycleCovariate.java index eaa98bb1d..5826263f9 100755 --- a/java/src/org/broadinstitute/sting/gatk/walkers/recalibration/CycleCovariate.java +++ b/java/src/org/broadinstitute/sting/gatk/walkers/recalibration/CycleCovariate.java @@ -143,10 +143,12 @@ public class CycleCovariate implements StandardCovariate { return getValue( read, offset ); // A recursive call } - // Differentiate between first and second of pair + // Differentiate between first and second of pair. // The sequencing machine cycle keeps incrementing for the second read in a pair. So it is possible for a read group - // to have an error affecting quality at a particular cycle on only second of pair reads. Therefore the cycle covariate must differentiate - // between first and second of pair reads. + // to have an error affecting quality at a particular cycle on the first of pair which carries over to the second of pair. + // Therefore the cycle covariate must differentiate between first and second of pair reads. + // This effect can not be corrected by pulling out the first of pair and second of pair flags into a separate covariate because + // the current sequential model would consider the effects independently instead of jointly. if( read.getReadPairedFlag() && read.getSecondOfPairFlag() ) { cycle *= -1; }