Transcript Effects ignored if more than 1 #1347
Comments
|
Thanks for reporting this, @majkiw! I'll take a look this morning. |
|
These quick tests pass diff --git a/adam-core/src/test/scala/org/bdgenomics/adam/converters/TranscriptEffectConverterSuite.scala b/adam-core/src/test/scala/org/bdgenomics/adam/converters/TranscriptEffectConverterSuite.scala
index c63029a0..5c7927cf 100644
--- a/adam-core/src/test/scala/org/bdgenomics/adam/converters/TranscriptEffectConverterSuite.scala
+++ b/adam-core/src/test/scala/org/bdgenomics/adam/converters/TranscriptEffectConverterSuite.scala
@@ -35,6 +35,7 @@ class TranscriptEffectConverterSuite extends ADAMFunSuite {
final val INVALID_NUMBER = "T|upstream_gene_variant||TAS1R3|ENSG00000169962|transcript|ENST00000339381.5|protein_coding|1/2|c.-485C>T|||4|1/42|not a number|"
final val INVALID_FRACTION = "T|upstream_gene_variant||TAS1R3|ENSG00000169962|transcript|ENST00000339381.5|protein_coding|not a number/2|c.-485C>T|||4|1/42|453|"
final val VALID = "T|upstream_gene_variant||TAS1R3|ENSG00000169962|transcript|ENST00000339381.5|protein_coding|1/2|c.-485C>T|||4|1/42|453|"
+ final val MULTIPLE = VALID + "," + VALID + "," + VALID
var variant: Variant = null
var variantContext: VariantContext = null
@@ -134,6 +135,33 @@ class TranscriptEffectConverterSuite extends ADAMFunSuite {
assert(te.getMessages.isEmpty)
}
+ test("parse VCF ANN attribute with multiple transcript effects") {
+ val ann = TranscriptEffectConverter.parseAnn(MULTIPLE, ValidationStringency.STRICT)
+ assert(ann.length == 3)
+
+ ann.foreach(te => {
+ assert(te.getAlternateAllele == "T")
+ assert(te.getEffects.contains("upstream_gene_variant"))
+ assert(te.getGeneName == "TAS1R3")
+ assert(te.getGeneId == "ENSG00000169962")
+ assert(te.getFeatureType == "transcript")
+ assert(te.getFeatureId == "ENST00000339381.5")
+ assert(te.getBiotype == "protein_coding")
+ assert(te.getRank == 1)
+ assert(te.getTotal == 2)
+ assert(te.getTranscriptHgvs == "c.-485C>T")
+ assert(te.getProteinHgvs == null)
+ assert(te.getCdnaPosition == null)
+ assert(te.getCdnaLength == null)
+ assert(te.getCdsPosition == 4)
+ assert(te.getCdsLength == null)
+ assert(te.getProteinPosition == 1)
+ assert(te.getProteinLength == 42)
+ assert(te.getDistance == 453)
+ assert(te.getMessages.isEmpty)
+ })
+ }
+
test("convert to transcript effect from null VCF ANN attribute in variant context") {
when(variantContext.getAttributeAsString("ANN", null)).thenReturn(null)
@@ -253,6 +281,42 @@ class TranscriptEffectConverterSuite extends ADAMFunSuite {
})
}
+ test("convert to transcript effect from VCF ANN attribute in variant context multiple effects same alt allele") {
+ variant = Variant.newBuilder()
+ .setAlternateAllele("T")
+ .build()
+
+ when(variantContext.getAttributeAsString("ANN", null)).thenReturn(MULTIPLE)
+
+ val transcriptEffectsOpt = TranscriptEffectConverter.convertToTranscriptEffects(variant, variantContext)
+ assert(transcriptEffectsOpt.isDefined)
+
+ transcriptEffectsOpt.foreach(transcriptEffects => {
+ assert(transcriptEffects.size === 3)
+ transcriptEffects.foreach(te => {
+ assert(te.getAlternateAllele == "T")
+ assert(te.getEffects.contains("upstream_gene_variant"))
+ assert(te.getGeneName == "TAS1R3")
+ assert(te.getGeneId == "ENSG00000169962")
+ assert(te.getFeatureType == "transcript")
+ assert(te.getFeatureId == "ENST00000339381.5")
+ assert(te.getBiotype == "protein_coding")
+ assert(te.getRank == 1)
+ assert(te.getTotal == 2)
+ assert(te.getTranscriptHgvs == "c.-485C>T")
+ assert(te.getProteinHgvs == null)
+ assert(te.getCdnaPosition == null)
+ assert(te.getCdnaLength == null)
+ assert(te.getCdsPosition == 4)
+ assert(te.getCdsLength == null)
+ assert(te.getProteinPosition == 1)
+ assert(te.getProteinLength == 42)
+ assert(te.getDistance == 453)
+ assert(te.getMessages.isEmpty)
+ })
+ })
+ }
+
test("convert transcript effect to VCF ANN attribute value") {
val te = TranscriptEffect.newBuilder()
.setAlternateAllele("T")The problem may be that in the tests I'm mocking htsjdk with reasonable behavior, which of course may not be a reasonable thing to do. |
|
Hey,
Notice extra brackets in the start/end of this string that break further parsing. So for lists we need |
|
If the ANN VCF header line is |
|
Thank you for picking this up @heuermh. |
Hi,
I was trying to run current
masteragainst our VCF files and I have noticed that whenever there is more than one value in transcript effects, the list is returned empty.BTW we really appreciate that you added support for transcript effects - this was the last missing thing we required to be able to use ADAM and you added it just 2 weeks after we started testing. Thank you :)
I am attaching example for completeness, but I don't think it's needed, because it happens on every row with multiple transcript effects.
This would return 1 transcript effect:
This would return no transcript effects:
Note: If no one is working on this I will probably look into this myself if I have time today or tomorrow.
The text was updated successfully, but these errors were encountered: