Coloc results using GCST003043-5 missing from certain genes now

Community Feedback Bug reports
1

Hello,

I believe there are some colocalization results missing from some genes using GCST003043/GCST003044/GCST003045. For example, I believe there should be colocalization results on this page: Open Targets Genetics. I’m fairly certain there use to be results here and now they are missing. If you download the v2d_coloc parquet files and look in the data, you can find these coloc results. I think I’ve seen this recently in other study-locus pages as well.

Here is some example code to reidentify the results that I think should be on this page. This is after downloading all the parquet files in Index of /pub/databases/opentargets/genetics/latest/v2d_coloc into a folder.

v2d_coloc ← arrow::open_dataset(path_to_above_folder)
v2d_coloc_schema ← v2d_coloc$schema
v2d_coloc_schema[[“right_pos”]] ← Field$create(“right_pos”, int64())
v2d_coloc_schema[[“left_pos”]] ← Field$create(“left_pos”, int64())
v2d_coloc ← arrow::open_dataset(path_to_above_folder, schema = v2d_coloc_schema)

Liu.coloc.eQTL.H4_8.ADCY3 ← data.frame(v2d_coloc %>%
filter (
left_study == “GCST003043” & coloc_h4 > .8 & right_type == “eqtl” & left_chrom == “2” & left_pos > 24819169 & left_pos < 24920237
))

Thanks!

2

Hi Michael,

Is the coloc result you’re looking for present in the parquet file?

Does the parquet file have a n_vars column? This indicates how many variants overlapped between the QTL and the GWAS signal. I seem to recall there being a filter for COLOC results with n_vars <200. Not sure if it was removed yet.

Best wishes,
Xiangyu