Split studies and summary statistics for PMID:30104761

[eks-ebi]

This publication was curated some time ago, and split into 4 studies because associations were only reported in the paper for these 4 studies. However, summary statistics were later obtained manually from Open Targets and added.

Three studies correctly have only the one set of summary statistics matching the trait analysed in that study:

• GCST008371 = thyroid cancer
• GCST008372 = colorectal cancer
• GCST008373 = glaucoma

However, for the remaining study (GCST008370 = coronary artery disease), the FTP folder contains up to 1,403 summary statistics files for many different traits which were also analysed as part of the project. The plan was to split these into separate entries once the deposition app was in production, but this has not yet been done.

Tasks:

• [ ] There is no submission for this publication, so in order to split the studies, the data will first need to be extracted and recurated into a submission template.
• [ ] Upload new metadata
• [ ] Move summary statistics into the correct new folders

[earlEBI]

The original GCSTs will also need to be kept which will require developer time well before a data release is due to start.

[earlEBI]

@ljwh2 I don't think I can find the key anywhere in the paper that explains what each sumstats filename relates to... I found this file online. Do you think I can use this for mapping and for sample case control numbers? https://github.com/opentargets/genetics-sumstat-data/blob/master/extras/prepare_uk_biobank_gwas_catalog/SAIGE/1_stream_to_gcs/manifest/SAIGE_UKBiobank_URL_Nov2017_v3.tsv

Our sumstats files: http://ftp.ebi.ac.uk/pub/databases/gwas/summary_statistics/GCST008001-GCST009000/GCST008370/?C=N;O=D

[earlEBI]

PMID is currently closed for submission as it was previously curated through the Curation App. Status needs to be changed to 'open for submission' well before a DR starts, so that I can submit the new studies through the submission app.

[earlEBI]

To do:

• Move files to Globus - done
• Submit - done
• Bypass sumstats validation - done
• Change 4 GCSTs to original GCSTs - done
• Change yaml types to pre-GWAS-SSF - created ticket: https://app.zenhub.com/workspaces/gwas-59df823c4a6feb3786810391/issues/gh/ebispot/goci/1387
• rename main sumstats files in original GCST folders to GCST format - to do

I think I'll keep all the original files (.gz and .tbi) in GCST008370 directory for provenance. They will remain with raw names to differentiate.

[earlEBI]

This ticket can be closed @ljwh2

[ljwh2]

We previously agreed not to reformat these files since the effect allele was not clear, but the author has now confirmed (via gwas-info) that ALT=effect so we could format them to GWAS-SSF and harmonise, re-opening the ticket