Distance-based clustering in gwas-catalog data

Community Feedback
1

Dear Team,

I am not so clear about distance-based clustering in gwas-catalog data, can you give me some further explanation?

    • Identify studies with an abnormally large number of reported loci compared to number of loci after distance-based clustering (±500kb). Criteria for “abnormal” are set to number of loci > 10, and decrease in loci count after clustering of >10%. For “abnormal” studies (N≈120), apply distance based clustering.

:speech_balloon:

Thanks.

Shicheng

2

Hi Shicheng,

GWAS Catalog can contain none independent loci where summary statistics are bulk imported from supplementary tables. To correct this, the distance-based clustering step is used to remove redundant variants.

Xiangyu

1 Like
3

Thank you @Xiangyu. I’m going to mark this as resolved for now. @Shicheng_Guo please let us know if you have any further questions

4

Thank you Xiangyu and Kirill, Can you share me the script how it was performed? or Is it shared in the github page?
Thanks
Shicheng

5

Hi Xiangyu and Kirill,

Is there any internal discussion about my previous poster?

Thanks.

Shicheng

@Kirill_Tsukanov @Xiangyu

6

I’ll let @Xiangyu get back to you about where to find the relevant code. Unmarking as resolved for now.

I’ll get back to you separately on the other post about mendelian randomisation.

7

Hi Shicheng,

Yes, the distance based clumping is a part of the v2d processing pipeline, and the script for it is here.

Best wishes,
Xiangyu

1 Like