Completes #5260 on GitHub. (#5464)

* Changed SelectVariants so that it van handle multiple rsIDs separated by ';' in a VCF file. * The tests testKeepSelectionIDLiteral and testKeepSelectionIDFromFile broke due to the change in the test file complexExample1.vcf. I modified the file testSelectVariants_KeepSelectionID.vcf appropriately so that the tests pass as they should. * Changes due to review by David Benjamin and Phil Shapiro. * Made the code in VariantIDsVariantFilter's test function more concise.
broadinstitute · Nov 30, 2018 · fb01bf3 · fb01bf3
1 parent d95ceec
commit fb01bf3
Show file tree

Hide file tree

Showing 3 changed files with 5 additions and 1 deletion.
diff --git a/src/main/java/org/broadinstitute/hellbender/engine/filters/VariantIDsVariantFilter.java b/src/main/java/org/broadinstitute/hellbender/engine/filters/VariantIDsVariantFilter.java
@@ -5,6 +5,8 @@
 
 import java.util.Set;
 import java.util.LinkedHashSet;
+import java.util.*;
+import java.util.stream.*;
 
 /**
  * Keep only variants with any of these IDs.
@@ -22,6 +24,6 @@ public VariantIDsVariantFilter(Set<String> keepIDs) {
 
     @Override
     public boolean test(final VariantContext vc) {
-        return includeIDs.contains(vc.getID());
+        return Arrays.stream(vc.getID().split(";")).anyMatch(includeIDs::contains);
     }
 }
diff --git a/...g/broadinstitute/hellbender/tools/walkers/variantutils/SelectVariants/complexExample1.vcf b/...g/broadinstitute/hellbender/tools/walkers/variantutils/SelectVariants/complexExample1.vcf
@@ -29,5 +29,6 @@
 1	10044557	.	C	T	62.22	.	AF=0.50;AlleleBalance=0.68;DoC=31;HomopolymerRun=16;MAPQ0=0;NS=1;RMSMAPQ=85.34;SB=-0.00;SpanningDeletions=0	GT:DP:GQ	./.	./.	0/1:31:15
 1	10045603	.	AAAA	A	40.26	PASS	AF=0.50	GT:DP:GQ	0/1:15:8	0/0:15:8	1/1:15:8
 1	10045604	.	A	ACAT	41.26	PASS	AF=0.50	GT:DP:GQ	1/1:14:7	0/0:15:8	1/0:15:8
+1	10046982	testid0;testid1	C	T	32.01	PASS	AF=0.40;AlleleBalance=0.37;DoC=21;HomopolymerRun=0;MAPQ0=8;NS=1;RMSMAPQ=12.31;SB=-4.2;SpanningDeletions=0	GT:DP:GQ	1/0:3:49	./.	./.
 1	10048142	.	A	G	126.81	foo	AF=1.00;DoC=36;HomopolymerRun=4;MAPQ0=0;NS=1;RMSMAPQ=82.11;SB=-85.45;SpanningDeletions=0	GT:DP:GQ	0/1:15:8	0/0:15:8	1/1:15:8
 1	10048580	.	T	A	72.22	bar;baz	AF=0.50;AlleleBalance=0.69;DoC=33;HomopolymerRun=6;MAPQ0=0;NS=1;RMSMAPQ=84.45;SB=-27.41;SpanningDeletions=0	GT:DP:GQ	0/1:15:8	0/0:15:8	1/1:15:8
diff --git a/...tools/walkers/variantutils/SelectVariants/expected/testSelectVariants_KeepSelectionID.vcf b/...tools/walkers/variantutils/SelectVariants/expected/testSelectVariants_KeepSelectionID.vcf
@@ -23,3 +23,4 @@
 ##source=ArbitrarySource
 #CHROM	POS	ID	REF	ALT	QUAL	FILTER	INFO	FORMAT	NA00001	NA00002	NA00003
 1	10001292	testid1	G	A	12.22	PASS	AF=0.50;AlleleBalance=0.57;DoC=23;HomopolymerRun=0;MAPQ0=8;NS=1;RMSMAPQ=46.31;SB=-24.92;SpanningDeletions=0	GT:DP:GQ	1/0:23:19	./.	./.
+1	10046982	testid0;testid1	C	T	32.01	PASS	AF=0.40;AlleleBalance=0.37;DoC=21;HomopolymerRun=0;MAPQ0=8;NS=1;RMSMAPQ=12.31;SB=-4.2;SpanningDeletions=0	GT:DP:GQ	1/0:3:49	./.	./.