Phenotype data from UKBB

[jackosullivanoxford]

Check phenotype data we extracted from the UKBB, was it self-reported and ICD coding? Consider a sensitivity analysis with only ICD coding phenotypes. We could do three analyses: 1. Self-reported and ICD-coding, 2. ICD coding only, 3. Self-reported only

[jackosullivanoxford]

To do this:

1. Find the phenotype file we originally created, and, more importantly, find the script we used to generate this
2. Go to UKBB data on sherlock and re-create phenotype file

[jackosullivanoxford]

I have an R script (Create_phenotype_file.R # ~/../2nd_run_through ), but I still need to include self-reported ischemic strokes

Access self-reported codes via UKBB datashow case: http://biobank.ctsu.ox.ac.uk/crystal/field.cgi?id=20002

Note that only 28 participants self-reported as having ischemic stroke, but 2168 self-reported TIA. Still need to include these in phenotype file

[jackosullivanoxford]

As per Meta-stroke - look at algorithmically defined strokes: https://biobank.ctsu.ox.ac.uk/crystal/label.cgi?id=43