how to find mapping of arrays

[Jiagengjie]

For ukbREST query, we need to specify field, instance, and arrays. But how would I know that asthma iss recorded as the first array of field 20002, i.e., 20002_0_0, while hypertension is the second, 20002_0_1? Basically, where can I find the information about which diseases the subsequent arrays map to, like 20002_0_2, 20002_0_3, etc? The same questions apply to any fields that have multiple arrays. thanks.

[Milton Pividori]

Hi Gengjie,

Thank you for your question.

ukbREST makes this process easy by not requiring you to know which array contains the disease code. The structure of UK Biobank allows to specify the asthma code (1111) in any of the 34 arrays of data-field 20002 (or any "categorical multiple" data type) as far as I know. So in the standard way of processing this data, you would need to obtain all arrays for the field and check one by one to see if the code 1111 is present. You don't need to do this with ukbREST, since it's already handled. Take a look at the Github page (https://github.com/hakyimlab/ukbrest/), you need to use a YAML file for this (and this is, in fact, the recommended way, since you can then share this YAML file with others or provide it as part of your manuscript).

Regards,