Mutect2 WDL and GetSampleName can handle multiple sample names in BAM headers

scripts/mutect2_wdl/mutect2.wdl

@@ -442,22 +442,20 @@ task M2 {
         touch bamout.bam
         touch f1r2.tar.gz
         touch dataset.txt
-        echo "" > normal_name.txt
-
-        gatk --java-options "-Xmx~{command_mem}m" GetSampleName -R ~{ref_fasta} -I ~{tumor_reads} -O tumor_name.txt -encode \
-        ~{"--gcs-project-for-requester-pays " + gcs_project_for_requester_pays}
-        tumor_command_line="-I ~{tumor_reads} -tumor `cat tumor_name.txt`"
 
         if [[ ! -z "~{normal_reads}" ]]; then
-            gatk --java-options "-Xmx~{command_mem}m" GetSampleName -R ~{ref_fasta} -I ~{normal_reads} -O normal_name.txt -encode \
+            gatk --java-options "-Xmx~{command_mem}m" GetSampleName -R ~{ref_fasta} -I ~{normal_reads} -O normal_names.txt -encode \
             ~{"--gcs-project-for-requester-pays " + gcs_project_for_requester_pays}
-            normal_command_line="-I ~{normal_reads} -normal `cat normal_name.txt`"
+            # add "-normal " to the start of each line and " " to the end, then remove newlines
+            # to get -normal sample1 -normal sample2 etc
+            normal_sample_line=`awk '{ print "-normal", $0 }' normal_names.txt | tr '\n' ' '`
         fi
 
         gatk --java-options "-Xmx~{command_mem}m" Mutect2 \
             -R ~{ref_fasta} \
-            $tumor_command_line \
-            $normal_command_line \
+            -I ~{tumor_reads} \
+            ~{"-I " + normal_reads} \
+            $normal_sample_line \
             ~{"--germline-resource " + gnomad} \
             ~{"-pon " + pon} \
             ~{"-L " + intervals} \
@@ -513,8 +511,6 @@ task M2 {
         File unfiltered_vcf = "~{output_vcf}"
         File unfiltered_vcf_idx = "~{output_vcf_idx}"
         File output_bamOut = "bamout.bam"
-        String tumor_sample = read_string("tumor_name.txt")
-        String normal_sample = read_string("normal_name.txt")
         File stats = "~{output_stats}"
         File f1r2_counts = "f1r2.tar.gz"
         Array[File] tumor_pileups = glob("*tumor-pileups.table")

src/main/java/org/broadinstitute/hellbender/tools/GetSampleName.java

@@ -86,18 +86,16 @@ public void onTraversalStart() {
         }
 
         final List<String> sampleNames = getHeaderForReads().getReadGroups().stream().map(s -> s.getSample()).distinct().collect(Collectors.toList());
-        if (sampleNames.size() > 1) {
-            throw new UserException.BadInput("The given input bam has more than one unique sample name: " + StringUtils.join(sampleNames, ", "));
-        }
 
         if (sampleNames.size() == 0) {
             throw new UserException.BadInput("The given bam input has no sample names.");
         }
 
         try (final OutputStreamWriter fileWriter = new OutputStreamWriter(outputSampleNameFile.getOutputStream())) {
-            final String rawSample = sampleNames.get(0);
-            final String outputSample = urlEncode ? IOUtils.urlEncode(rawSample) : rawSample;
-            fileWriter.write(outputSample);
+            final String outputSamplesOnSeparateLines = sampleNames.stream()
+                    .map(rawSample -> urlEncode ? IOUtils.urlEncode(rawSample) : rawSample)
+                    .collect(Collectors.joining("\n"));
+            fileWriter.write(outputSamplesOnSeparateLines);
         } catch (final IOException ioe) {
             throw new UserException(String.format("Could not write to output file %s.", outputSampleNameFile), ioe);
         }

src/test/java/org/broadinstitute/hellbender/tools/GetSampleNameIntegrationTest.java

@@ -47,13 +47,16 @@ public void testUrlEncoding() throws IOException {
         Assert.assertTrue(Files.readAllLines(outputFile.toPath()).stream().filter(n -> n.equals("Hi%2CMom%21")).count() == 1);
     }
 
-    @Test(expectedExceptions = UserException.class)
-    public void testMultiSampleBam() {
+    @Test
+    public void testMultiSampleBam() throws IOException {
         final File outputFile = createTempFile("get-sample-name-ms", ".txt");
         final String[] arguments = {
                 "-" + StandardArgumentDefinitions.INPUT_SHORT_NAME, BAD_MULTI_SAMPLE_BAM_FILE.getAbsolutePath(),
                 "-" + StandardArgumentDefinitions.OUTPUT_SHORT_NAME, outputFile.getAbsolutePath()
         };
         runCommandLine(arguments);
+        Assert.assertTrue(outputFile.exists());
+        Assert.assertTrue(outputFile.length() > 0);
+        Assert.assertTrue(Files.readAllLines(outputFile.toPath()).stream().count() == 3);
     }
 }