Hi - should work fine, it is not dependent on RStudio at all - just the way I use it! Any issues or suggestions let me know (via GitHub ideally!)

🤞 for effective sample size

Worth following hugeraproject.bsky.social if you are also interesting - they are hosting workshops and seminars to further the conversation!

I find these cases fascinating, where absence of genetic correlation does not imply absence of shared causal variants!

Yeah... at the least could use {renv} to make the same package versions available etc

But my code doesn't try to emulate the real data structures particularly - just col names and variable types, so that at least the initial test can be done there, then tested properly online

This is really cool Michel. I have been making something for Our Future Health, because of their "airlock" it's much more tedious to test code. Whereas UK Biobank it is very easy (and pretty cheap) to test code on the real data (for most cases anyway - no doubt there are cases where this isn't true)

Can be very effective in talks/slides, but for papers the default "white" background is ubiquitous, even in online-only publications!