search

opentargets / issues

Issue tracker for Open Targets Platform and Open Targets Genetics Portal
https://platform.opentargets.org https://genetics.opentargets.org
Apache License 2.0
12 stars 2 forks source link

Beta directions for immune related traits are incorrect #2077

Closed edm1 closed 2 years ago

edm1 commented 5 years ago

Beta directions for a number of immune traits including de Lange (GCST004131) or Liu (GCST003043) are incorrect

As can be seen in the GWAS Coloc table here: https://dev--genetics-app.netlify.com/study-locus/GCST003044/2_233252937_T_C

edm1 commented 5 years ago

These studies have incorrect beta directions and will be fixed in post-production:

study_id pmid pub_date pub_journal pub_title pub_author trait_reported
GCST004131 PMID:28067908 2017-01-09 Nat Genet Genome-wide association study implicates immune activation of multiple integrin genes in inflammatory bowel disease. de Lange KM Inflammatory bowel disease
GCST004132 PMID:28067908 2017-01-09 Nat Genet Genome-wide association study implicates immune activation of multiple integrin genes in inflammatory bowel disease. de Lange KM Crohn's disease
GCST004133 PMID:28067908 2017-01-09 Nat Genet Genome-wide association study implicates immune activation of multiple integrin genes in inflammatory bowel disease. de Lange KM Ulcerative colitis
GCST001725 PMID:23128233 2012-11-01 Nature Host-microbe interactions have shaped the genetic architecture of inflammatory bowel disease. Jostins L Inflammatory bowel disease
GCST001728 PMID:23128233 2012-11-01 Nature Host-microbe interactions have shaped the genetic architecture of inflammatory bowel disease. Jostins L Ulcerative colitis
GCST001729 PMID:23128233 2012-11-01 Nature Host-microbe interactions have shaped the genetic architecture of inflammatory bowel disease. Jostins L Crohn's disease
GCST000964 PMID:21297633 2011-02-06 Nat Genet Meta-analysis identifies 29 additional ulcerative colitis risk loci, increasing the number of confirmed associations to 47. Anderson CA Ulcerative colitis

Study IDs

GCST004131
GCST004132
GCST004133
GCST001725
GCST001728
GCST001729
GCST000964
edm1 commented 5 years ago

@mkarmona Here are the locations of the updated files

Top loci

Credible sets

Sumstats

Coloc

edm1 commented 5 years ago

This is fixed in the current version (post-production), but is not fixed at source. To fix at source requires:

ktsirigos commented 3 years ago

Hi @Jeremy37, has this been fixed?

buniello commented 2 years ago

@Jeremy37 @ktsirigos - this issue has been fixed at source (GWAS Cat harmonised summary statistics) back in 2019.

I have noticed, however, that in OTAR Gen betas from De Lange 2017 (GCST004131, GCST004132, GCST004133) are reported with opposite direction from what is made available in the Catalog harmonised files. I know you also recently received a user query reporting this discrepancy between OTAr and the Catalog. Please note the effect alleles reported in OTAR are in line with the GWAS harm file, only betas are inverted at your end.

Example from the user: Taking an IRF5 variant 7_128933913_G_A as an example, we see that the effect allele in the GWAS Catalog summary stats is A and the beta is 0.0995. However, in Open Targets the effect allele is A and the beta is -0.0995. https://genetics.opentargets.org/variant/7_128933913_G_A.

Variants I have checked to confirm behaviour: 10_99527227_C_T (UC), 10_99558359_G_A (UC), 10_62748961_A_G (CD),, 10_79306820_G_A (CD), 10_62685804_A_G (IBD), 10_62756305_G_T (IBD), 10_62657529_C_T (IBD).

I hope this helps!

d0choa commented 2 years ago

@buniello when you have a few minutes, could you confirm genetics portal looks in line with GWAS Catalog?

buniello commented 2 years ago

@d0choa: yes, triaged on a number of variants from the three De Lange studies and genetics portal betas are now in line with GWAS Catalog harmonised summary statistics

d0choa commented 2 years ago

Thanks for you help @buniello!