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

Fix sample sort order bug in DepthOfCoverage 

Rare bug triggered by hash collision between sample names
 PT 66183936
This commit is contained in:
Phillip Dexheimer 2014-08-05 21:55:34 -04:00
parent 03e7ee6e9c
commit b0c026e671
3 changed files with 31 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
public/gatk-tools-public/src/main/java/org/broadinstitute/gatk/tools/walkers/coverage/DepthOfCoverage.java
View File

@ -1054,7 +1054,7 @@ class CoveragePartitioner {
private Map<DoCOutputType.Partition,List<String>> identifiersByType;
private Set<String> allIdentifiers;
public CoveragePartitioner(Collection<DoCOutputType.Partition> typesToUse, int start, int stop, int nBins) {
coverageProfiles = new HashMap<DoCOutputType.Partition,DepthOfCoverageStats>();
coverageProfiles = new TreeMap<DoCOutputType.Partition,DepthOfCoverageStats>();
identifiersByType = new HashMap<DoCOutputType.Partition,List<String>>();
types = typesToUse;
for ( DoCOutputType.Partition type : types ) {

5
public/gatk-tools-public/src/main/java/org/broadinstitute/gatk/tools/walkers/coverage/DepthOfCoverageStats.java
View File

@ -31,6 +31,7 @@ import org.broadinstitute.gatk.utils.exceptions.UserException;
import java.util.HashMap;
import java.util.Map;
import java.util.Set;
import java.util.TreeMap;
/**
* IF THERE IS NO JAVADOC RIGHT HERE, YELL AT chartl
@ -111,8 +112,8 @@ public class DepthOfCoverageStats {
public DepthOfCoverageStats(DepthOfCoverageStats cloneMe) {
this.binLeftEndpoints = cloneMe.binLeftEndpoints;
granularHistogramBySample = new HashMap<String,long[]>();
totalCoverages = new HashMap<String,Long>();
granularHistogramBySample = new TreeMap<String,long[]>();
totalCoverages = new TreeMap<String,Long>();
for ( String s : cloneMe.getAllSamples() ) {
granularHistogramBySample.put(s,new long[cloneMe.getHistograms().get(s).length]);
for ( int i = 0; i < granularHistogramBySample.get(s).length; i++ ) {

27
public/gatk-tools-public/src/test/java/org/broadinstitute/gatk/tools/walkers/coverage/DepthOfCoverageIntegrationTest.java
View File

@ -142,6 +142,33 @@ public class DepthOfCoverageIntegrationTest extends WalkerTest {
execute("testAdjacentIntervals", spec);
}
@Test
public void testSortOrder() {
// This test came from a user who discovered that the columns and data in the gene_summary file didn't align for the specific
// sample names in these files.
String[] intervals = {"1:1600000-1700000"};
String[] bams = {privateTestDir+"badHashName1.bam", privateTestDir+"badHashName2.bam"};
String cmd = buildRootCmd(b37KGReference, new ArrayList<String>(Arrays.asList(bams)), new ArrayList<String>(Arrays.asList(intervals))) +
" -geneList "+privateTestDir+"refGene_CDK11B.txt";
WalkerTestSpec spec = new WalkerTestSpec(cmd, 0, new ArrayList<String>());
File baseOutputFile = WalkerTest.createTempFile("depthofcoveragesortorder", ".tmp");
spec.setOutputFileLocation(baseOutputFile);
spec.addAuxFile("a148e50f9db207adfd5d5f0f29eb54d8", baseOutputFile);
spec.addAuxFile("7ccd5193a3c035d1cc856cbc89e3daf4", createTempFileFromBase(baseOutputFile.getAbsolutePath()+".sample_cumulative_coverage_counts"));
spec.addAuxFile("2efe59c20721ce61bc5b334a26d11720", createTempFileFromBase(baseOutputFile.getAbsolutePath()+".sample_cumulative_coverage_proportions"));
spec.addAuxFile("9194cec953e0fe0b84a681f9bb63ffbe", createTempFileFromBase(baseOutputFile.getAbsolutePath()+".sample_gene_summary"));
spec.addAuxFile("cf62d95ec1f459fbbe35370c3f0ca481", createTempFileFromBase(baseOutputFile.getAbsolutePath()+".sample_interval_statistics"));
spec.addAuxFile("b4fcb739b7f9e309e38a7d5e7e4ebb9f", createTempFileFromBase(baseOutputFile.getAbsolutePath()+".sample_interval_summary"));
spec.addAuxFile("6bf63f9c62071e850c6f0b6356fb63eb", createTempFileFromBase(baseOutputFile.getAbsolutePath()+".sample_statistics"));
spec.addAuxFile("e53e6a494bf1cf817762b74917c6f0c9", createTempFileFromBase(baseOutputFile.getAbsolutePath()+".sample_summary"));
execute("testSortOrder", spec);
}
public void testRefNHandling(boolean includeNs, final String md5) {
String command = "-R " + b37KGReference + " -L 20:26,319,565-26,319,575 -I " + validationDataLocation + "NA12878.HiSeq.WGS.bwa.cleaned.recal.hg19.20.bam -T DepthOfCoverage -baseCounts --omitIntervalStatistics --omitLocusTable --omitPerSampleStats -o %s";
if ( includeNs ) command += " --includeRefNSites";