No duplicates on GRCh38

bpblanken · bpblanken · commit b3708d40fdd4 · 2024-12-04T11:20:12.000-05:00
diff --git a/v03_pipeline/lib/reference_datasets/topmed.py b/v03_pipeline/lib/reference_datasets/topmed.py
@@ -15,7 +15,8 @@
 
 def get_ht(path: str, reference_genome: ReferenceGenome) -> hl.Table:
     ht = vcf_to_ht(path, reference_genome)
-    ht = ht.filter(ht.locus.position == hl.int(ht.info.OriginalStart))
+    if reference_genome == ReferenceGenome.GRCh37:
+        ht = ht.filter(ht.locus.position == hl.int(ht.info.OriginalStart))
     return ht.select(
         **{k: parse_nested_field(ht, v) for k, v in SELECT.items()},
     )

Original file line number	Diff line number	Diff line change
`@@ -15,7 +15,8 @@`
`15`	`15`
`16`	`16`	`def get_ht(path: str, reference_genome: ReferenceGenome) -> hl.Table:`
`17`	`17`	`ht = vcf_to_ht(path, reference_genome)`
`18`		`- ht = ht.filter(ht.locus.position == hl.int(ht.info.OriginalStart))`
	`18`	`+ if reference_genome == ReferenceGenome.GRCh37:`
	`19`	`+ ht = ht.filter(ht.locus.position == hl.int(ht.info.OriginalStart))`
`19`	`20`	`return ht.select(`
`20`	`21`	`**{k: parse_nested_field(ht, v) for k, v in SELECT.items()},`
`21`	`22`	`)`