zzh
/
gatk-3.8
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Fixed the RR bug I (knowingly) introduced last week: turns out we can't trust a context size's worth of data from the previous marking. I think Mauricio warned me about this but I forgot.

This commit is contained in:
Eric Banks 2012-10-22 11:48:34 -04:00
parent 9f2851d769
commit ccae6a5b92
1 changed files with 4 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
protected/java/src/org/broadinstitute/sting/gatk/walkers/compression/reducereads/SlidingWindow.java
View File

@ -276,7 +276,10 @@ public class SlidingWindow {
final int windowHeaderStartLocation = getStartLocation(windowHeader);
final int sizeOfMarkedRegion = stop - windowHeaderStartLocation + contextSize + 1;
final int lastPositionMarked = markedSites.updateRegion(windowHeaderStartLocation, sizeOfMarkedRegion);
// copy over as many bits as we can from the previous calculation. Note that we can't trust the
// last (contextSize - 1) worth of bits because we may not have actually looked at variant regions there.
final int lastPositionMarked = markedSites.updateRegion(windowHeaderStartLocation, sizeOfMarkedRegion) - contextSize - 1;
final int locationToProcess = Math.min(lastPositionMarked, stop - contextSize);
// update the iterator to the correct position