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

It turns out that it is legitimately possible for there to be reads that won't overlap within a target interval for cleaning. While we don't want to attempt cleaning, we also don't want to fail. 

git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@3682 348d0f76-0448-11de-a6fe-93d51630548a
This commit is contained in:
ebanks 2010-06-30 15:50:44 +00:00
parent ae33d8a2f2
commit 47c4a70ac1
1 changed files with 20 additions and 9 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

29
java/src/org/broadinstitute/sting/gatk/walkers/indels/IndelRealigner.java
View File

@ -264,18 +264,16 @@ public class IndelRealigner extends ReadWalker<Integer, Integer> {
if ( doNotTryToClean(read) ) {
readsNotToClean.add(read);
} else {
readsToClean.add(read, ref.getBases());
boolean wellCoveredInterval = readsToClean.add(read, ref.getBases());
if ( !wellCoveredInterval )
abortCleanForCurrentInterval();
// add the rods to the list of known variants
populateKnownIndels(metaDataTracker, ref);
}
if ( readsToClean.size() + readsNotToClean.size() >= MAX_READS ) {
// merge the two sets for emission
readsNotToClean.addAll(readsToClean.getReads());
emit(readsNotToClean);
readsToClean.clear();
readsNotToClean.clear();
currentInterval = intervals.hasNext() ? intervals.next() : null;
abortCleanForCurrentInterval();
}
}
else { // the read is past the current interval
@ -285,6 +283,15 @@ public class IndelRealigner extends ReadWalker<Integer, Integer> {
return 0;
}
private void abortCleanForCurrentInterval() {
// merge the two sets for emission
readsNotToClean.addAll(readsToClean.getReads());
emit(readsNotToClean);
readsToClean.clear();
readsNotToClean.clear();
currentInterval = intervals.hasNext() ? intervals.next() : null;
}
private boolean doNotTryToClean(SAMRecord read) {
return read.getReadUnmappedFlag() ||
read.getNotPrimaryAlignmentFlag() ||
@ -1232,7 +1239,9 @@ public class IndelRealigner extends ReadWalker<Integer, Integer> {
public ReadBin() { }
public void add(SAMRecord read, byte[] ref) {
// Return false if we can't process this read bin because the reads are not correctly overlapping.
// This can happen if e.g. there's a large known indel with no overlapping reads.
public boolean add(SAMRecord read, byte[] ref) {
reads.add(read);
// set up the reference
@ -1243,7 +1252,7 @@ public class IndelRealigner extends ReadWalker<Integer, Integer> {
long lastPosWithRefBase = loc.getStart() + reference.length -1;
int neededBases = (int)(read.getAlignmentEnd() - lastPosWithRefBase);
if ( neededBases > ref.length )
throw new StingException("Read " + read.getReadName() + " does not overlap the previous read in this interval; please ensure that you are using the same input bam that was used in the RealignerTargetCreator step");
return false;
if ( neededBases > 0 ) {
byte[] newReference = new byte[reference.length + neededBases];
System.arraycopy(reference, 0, newReference, 0, reference.length);
@ -1252,6 +1261,8 @@ public class IndelRealigner extends ReadWalker<Integer, Integer> {
loc = GenomeLocParser.createGenomeLoc(loc.getContigIndex(), loc.getStart(), loc.getStop()+neededBases);
}
}
return true;
}
public List<SAMRecord> getReads() { return reads; }