VariantFiltration: added arg to write custom mask filter description …

…in VCF header (#8831) Added a --mask-description argument to VariantFiltration to write a custom description for the mask filter in the VCF header
broadinstitute · Jul 15, 2024 · 78b8231 · 78b8231
1 parent f2a85b8
commit 78b8231
Show file tree

Hide file tree

Showing 2 changed files with 33 additions and 1 deletion.
diff --git a/src/main/java/org/broadinstitute/hellbender/tools/walkers/filters/VariantFiltration.java b/src/main/java/org/broadinstitute/hellbender/tools/walkers/filters/VariantFiltration.java
@@ -114,6 +114,7 @@ public final class VariantFiltration extends VariantWalker {
     public static final String CLUSTER_WINDOW_SIZE_LONG_NAME = "cluster-window-size";
     public static final String MASK_EXTENSION_LONG_NAME = "mask-extension";
     public static final String MASK_NAME_LONG_NAME = "mask-name";
+    public static final String MASK_DESCRIPTION_LONG_NAME = "mask-description";
     public static final String FILTER_NOT_IN_MASK_LONG_NAME = "filter-not-in-mask";
     public static final String MISSING_VAL_LONG_NAME = "missing-values-evaluate-as-failing";
     public static final String INVERT_LONG_NAME = "invert-filter-expression";
@@ -238,6 +239,14 @@ public final class VariantFiltration extends VariantWalker {
     @Argument(fullName=ALLELE_SPECIFIC_LONG_NAME, optional=true, doc="Set mask at the allele level. This option is not compatible with clustering.")
     public boolean applyForAllele = false;
 
+    /**
+     * If a mask interval list is provided, then set the description of the filter in the VCF header to this String.
+     * Note that if spaces are needed, then the entire description should be enclosed in quotes. Also note that if
+     * --filter-not-in-mask is used, the description should be adapted to reflect the reverse logic.
+     */
+    @Argument(fullName=MASK_DESCRIPTION_LONG_NAME, optional=true, doc="Description to add to the FILTER field in VCF header for the mask filter.")
+    public String maskDescription;
+
     // JEXL expressions for the filters
     private List<JexlVCMatchExp> filterExps;
     private List<JexlVCMatchExp> genotypeFilterExps;
@@ -305,7 +314,9 @@ private void initializeVcfWriter() {
             }
 
             if ( mask != null ) {
-                if (filterRecordsNotInMask) {
+                if (maskDescription != null) {
+                    hInfo.add(new VCFFilterHeaderLine(maskName, maskDescription));
+                } else if (filterRecordsNotInMask) {
                     hInfo.add(new VCFFilterHeaderLine(maskName, "Doesn't overlap a user-input mask"));
                 } else {
                     hInfo.add(new VCFFilterHeaderLine(maskName, "Overlaps a user-input mask"));
@@ -331,6 +342,9 @@ public void onTraversalStart() {
         if (filterRecordsNotInMask && mask == null) {
             throw new CommandLineException.BadArgumentValue(FILTER_NOT_IN_MASK_LONG_NAME, "argument not allowed if mask argument is not provided");
         }
+        if (maskDescription != null && mask == null) {
+            throw new CommandLineException.BadArgumentValue(MASK_DESCRIPTION_LONG_NAME, "argument not allowed if mask argument is not provided");
+        }
         filterExps = VariantContextUtils.initializeMatchExps(filterNames, filterExpressions);
         genotypeFilterExps = VariantContextUtils.initializeMatchExps(genotypeFilterNames, genotypeFilterExpressions);
         howToTreatMissingValues = failMissingValues ? JexlMissingValueTreatment.TREAT_AS_MATCH : JexlMissingValueTreatment.TREAT_AS_MISMATCH;

diff --git a/...org/broadinstitute/hellbender/tools/walkers/filters/VariantFiltrationIntegrationTest.java b/...org/broadinstitute/hellbender/tools/walkers/filters/VariantFiltrationIntegrationTest.java
@@ -1,12 +1,14 @@
 package org.broadinstitute.hellbender.tools.walkers.filters;
 
 import htsjdk.variant.variantcontext.VariantContext;
+import htsjdk.variant.vcf.VCFHeader;
 import org.broadinstitute.hellbender.CommandLineProgramTest;
 import org.broadinstitute.hellbender.cmdline.StandardArgumentDefinitions;
 import org.broadinstitute.hellbender.engine.FeatureDataSource;
 import org.broadinstitute.hellbender.exceptions.UserException;
 import org.broadinstitute.hellbender.testutils.ArgumentsBuilder;
 import org.broadinstitute.hellbender.testutils.IntegrationTestSpec;
+import org.broadinstitute.hellbender.testutils.VariantContextTestUtils;
 import org.testng.Assert;
 import org.testng.annotations.DataProvider;
 import org.testng.annotations.Test;
@@ -91,6 +93,22 @@ public void testMaskReversed() throws IOException {
         spec.executeTest("testMaskReversed", this);
     }
 
+    @Test
+    public void testMaskDescription() throws IOException {
+        final File output = createTempFile("testHeader", ".vcf");
+
+        final ArgumentsBuilder args = new ArgumentsBuilder();
+        args.add("V", getTestFile("vcfexample2.vcf"))
+                .add("mask-name", "foo")
+                .add("mask-description", "description")
+                .add("mask:BED", getToolTestDataDir() + "goodMask.bed");
+        args.addOutput(output);
+
+        runCommandLine(args);
+        VCFHeader header = VariantContextTestUtils.getVCFHeader(output.getAbsolutePath());
+        header.getFilterLines().stream().filter(f -> f.getID().equals("foo")).forEach(f -> Assert.assertEquals(f.getDescription(), "description"));
+    }
+
     @Test
     public void testIllegalFilterName() throws IOException {
         final IntegrationTestSpec spec = new IntegrationTestSpec(