Add support for Sam Header Readgroup Barcode field (#1210)

* adding support in SAMReadGroupRecord for the BC attribute
samtools · Jan 29, 2019 · 5ef9223 · 5ef9223
1 parent 3c48018
commit 5ef9223
Show file tree

Hide file tree

Showing 3 changed files with 68 additions and 9 deletions.
diff --git a/src/main/java/htsjdk/samtools/SAMReadGroupRecord.java b/src/main/java/htsjdk/samtools/SAMReadGroupRecord.java
@@ -25,12 +25,9 @@
 
 
 import htsjdk.samtools.util.Iso8601Date;
+import htsjdk.samtools.util.SamConstants;
 
-import java.util.Arrays;
-import java.util.Date;
-import java.util.HashSet;
-import java.util.Map;
-import java.util.Set;
+import java.util.*;
 
 /**
  * Header information about a read group.
@@ -51,6 +48,8 @@ public class SAMReadGroupRecord extends AbstractSAMHeaderRecord
     public static final String PLATFORM_MODEL_TAG = "PM";
     public static final String PLATFORM_UNIT_TAG = "PU";
     public static final String READ_GROUP_SAMPLE_TAG = "SM";
+    public static final String BARCODE_TAG = "BC";
+
 
     /* Platform values for the @RG-PL tag */
     public enum PlatformValue {
@@ -63,7 +62,7 @@ public enum PlatformValue {
             new HashSet<String>(Arrays.asList(READ_GROUP_ID_TAG, SEQUENCING_CENTER_TAG, DESCRIPTION_TAG,
                     DATE_RUN_PRODUCED_TAG, FLOW_ORDER_TAG, KEY_SEQUENCE_TAG, LIBRARY_TAG,
                     PROGRAM_GROUP_TAG, PREDICTED_MEDIAN_INSERT_SIZE_TAG, PLATFORM_TAG, PLATFORM_MODEL_TAG,
-                    PLATFORM_UNIT_TAG, READ_GROUP_SAMPLE_TAG));
+                    PLATFORM_UNIT_TAG, READ_GROUP_SAMPLE_TAG, BARCODE_TAG));
 
     public SAMReadGroupRecord(final String id) { mReadGroupId = id; }
 
@@ -90,6 +89,36 @@ public SAMReadGroupRecord(final String id, final SAMReadGroupRecord srcProgramRe
     public String getPlatform() { return getAttribute(PLATFORM_TAG); }
     public void setPlatform(final String platform) { setAttribute(PLATFORM_TAG, platform); }
 
+    /**
+     * @return the List of barcodes associated with this read group or null
+     */
+    public List<String> getBarcodes() {
+        final String barcodeString = getAttribute(BARCODE_TAG);
+        if (barcodeString == null) {
+            return null;
+        } else if (barcodeString.isEmpty()) {
+            return Collections.emptyList();
+        } else {
+            return Arrays.asList(barcodeString.split(SamConstants.BARCODE_SEQUENCE_DELIMITER));
+        }
+    }
+
+    /**
+     * Set the barcodes associated with this ReadGroup.
+     * Note that an input of null results in unsetting the attribute while an empty list is set as a tag with an empty value.
+     * @param barcodes a list of barcodes to associate with this read group
+     */
+    public void setBarcodes(final List<String> barcodes) {
+        if (barcodes == null) {
+            setAttribute(BARCODE_TAG, null);
+        } else {
+            if (barcodes.stream().anyMatch(String::isEmpty)) {
+                throw new IllegalArgumentException("A barcode must not be an empty String");
+            }
+           setAttribute(BARCODE_TAG, String.join(SamConstants.BARCODE_SEQUENCE_DELIMITER, barcodes));
+        }
+    }
+
     public Date getRunDate() {
         final String dt = getAttribute(DATE_RUN_PRODUCED_TAG);
         if (dt == null) return null;

diff --git a/src/main/java/htsjdk/samtools/util/SamConstants.java b/src/main/java/htsjdk/samtools/util/SamConstants.java
@@ -30,9 +30,16 @@
  */
  public final class SamConstants {
 
-    //No need to instantiate this class since all the onstants should be static
+    //No need to instantiate this class since all the constants should be static
     private SamConstants(){};
 
-    final static String BARCODE_SEQUENCE_DELIMITER = "-";
-    final static String BARCODE_QUALITY_DELIMITER = " ";
+    /**
+     * The recommended separator to use when specifying multiple barcodes together in the same tag.
+     */
+    public static final String BARCODE_SEQUENCE_DELIMITER = "-";
+
+    /**
+     * The recommend separator to use when specifying multiple barcode quality scores together in the same tag.
+     */
+    public static final String BARCODE_QUALITY_DELIMITER = " ";
 }
diff --git a/src/test/java/htsjdk/samtools/SAMReadGroupRecordTest.java b/src/test/java/htsjdk/samtools/SAMReadGroupRecordTest.java
@@ -29,7 +29,10 @@
 import org.testng.annotations.DataProvider;
 import org.testng.annotations.Test;
 
+import java.util.Arrays;
+import java.util.Collections;
 import java.util.Date;
+import java.util.List;
 import java.util.function.BiConsumer;
 import java.util.function.Function;
 
@@ -145,4 +148,24 @@ public void testEqualsAndHashcode(final SAMReadGroupRecord rg, final Object othe
         }
     }
 
+    @DataProvider
+    public Object[][] getBarcodes() {
+        return new Object[][] {
+                {null, null},
+                {Collections.emptyList(), ""},
+                {Collections.singletonList("aa"), "aa"},
+                {Arrays.asList("aa", "ac"), "aa-ac"},
+                {Arrays.asList("aa", "ca", "gg"), "aa-ca-gg"}
+        };
+    }
+
+    @Test(dataProvider = "getBarcodes")
+    public void testGetAndSetBarcodes(List<String> barcodes, String encoded){
+        final SAMReadGroupRecord readGroup = new SAMReadGroupRecord("ReadGroup");
+        Assert.assertNull(readGroup.getBarcodes());
+        Assert.assertNull(readGroup.getAttribute(SAMReadGroupRecord.BARCODE_TAG));
+        readGroup.setBarcodes(barcodes);
+        Assert.assertEquals(readGroup.getBarcodes(), barcodes);
+        Assert.assertEquals(readGroup.getAttribute(SAMReadGroupRecord.BARCODE_TAG), encoded);
+    }
 }