In Joan Didion’s essay collection Slouching Towards Bethlehem she describes her notebooks as full of “bits of the mind’s string too short to use.” Here is a list of lists of bits of string.
Note: I do not require that whatever I put here is correct nor do I require that it should be useful. I aim for eventual correctness and potential usefulness with no guarantees. This is to lower the barrier for writing.
(§) An old blog post by Rafa Irizarry tells us that lowering the GWAS threshold would save millions of dollars. The point: if you accept a larger false discovery rate you get more statistical power. It is clear that you can trade type I error for type II error and for expensive genomics procedures perhaps the odd “false discovery” is cheaper than sequencing more samples. Don’t think that this particular concern is too important almost 10 years later, but it is an interesting question in general: do you want to spend money or shoeleather?
this file last touched 2026.02.06