1/1
UKBiobank depression PRS weights
This folder contains the clumped summary statistics (in daner format) of GWAS performed on different definitions of depression in UKBiobank described in Cai N., et al. Minimal phenotyping yields genome-wide association signals of low specificity for major depression. Nature Genetics (2020) doi:10.1038/s41588-020-0594-5.
These files are used for assessing polygenic risk score (PRS) prediction of depression status in individual cohorts of Major Depressive Disorder Psychiatric Genomics Consortium from the PGC29 release, as described in the paper. GWAS summary statistics from this study, accessible through https://doi.org/10.6084/m9.figshare.11733753.v1, are first pruned for SNPs present in the PGC29 cohorts, with their effect alleles also aligned with those specified in PGC29 cohorts.
They then undergo LD-clumping using the clump_nav3 function in the Ricopilli pipeline, with the following parameters:
clump_nav3 --noindel \
--pfile $pheno_clumpOut.gz \
--hq_f 0.01 \
--hq_i 0.6 \
--outname $pheno_clumpOut \
--clu_p1 1.0 \
--clu_p2 1.0 \
--clu_window 500 \
--clu_r2 0.1 \
--refdir $refdir
--popname eur \
--force1
They are then used in PRS prediction of MDD status in PGC29 cohorts using the my_postimp_navi function in Ricopilli.
Please note that these files do not contain SNPs filtered based on P value thresholds; they contain SNPs across the whole range of association P values in their respective GWAS, and P value thresholds were only imposed when calculating PRS prediction statistics.
====
The phenotypes are described briefly below. For more details please see description in the paper and its supplemental methods. All phenotypes are case-control phenotypes (coded cases = 1, controls = 0), defined using criteria as described. Cases are those who have answered the relevant questions in UKBiobank and fulfil case criteria, controls are those who have answered the same questions but did not fulfil the case criteria.
Please note there are overlaps between cases and controls between phenotypes, these are described in the paper and its supplemental methods.
1. LifetimeMDD - Lifetime MDD derived with DSM-V symptom and impairment criteria using the PHQ-9 questionnaire in the Online Mental Health follow-up in UKBiobank
2. MDDRecur - Lifetime MDD with recurrence derived with DSM-V symptom and impairment criteria using the PHQ-9 questionnaire in the Online Mental Health follow-up in UKBiobank, with additional criteria on recurrence
3. GPpsy - Having gone to a General Practitioner (GP) for nerves, anxiety, tension or depression, answered in the touchscreen interview in UKBiobank. Please note this is the most similar phenotype to "Broad Depression" introduced in Howard et al 2018 Nature Communications (doi:10.1038/s41467-018-03819-3)
4. Psypsy - Having gone to a psychiatrist for nerves, anxiety, tension or depression, answered in the touchscreen interview in UKBiobank
5. DepAll - Having either of the two cardinal symptoms for MDD (low mood or anhedonia) for more than two weeks in addition to having gone to either the GP or psychiatrist for nerves, anxiety, tension or depression, answered in the touchscreen interview in UKBiobank. Please note phenotype is first introduced as "Probable Depression" in Smith et al 2013 PloS One (doi:10.1371/journal.pone.0075362.s001), and the most similar phenotype to "Probable Depression" introduced in Howard et al 2018 Nature Communications (doi:10.1038/s41467-018-03819-3)
6. GPNoDep - Having gone to the GP for nerves, anxiety, tension or depression, but did not report having cardinal symptoms of MDD in the touchscreen interview in UKBiobank.
7. SelfRepDep - Self-reported depression or depression symptoms described to a trained nurse during verbal interview of medical conditions, classified under "Non-cancer illness" in the UKBiobank dataset.
