News for hackers. Stuff that matters.
Re:Vitamin D uncorrelated with Covid severity in those of Eu... by Gatsky on Wednesday January 20, 2021 @06:21PM (#25853233)
Beyond these methodological questions, I think that the Mendelian Randomisation approach works best for situations where the effect of some exposure is chronic eg LDL and atherosclerosis, particularly where the disease process may begin in childhood or adolescence.

For an acute infection I think it is a blunt instrument at best, as there is going to be a lot more noise in the inferred target variable.

Re:Vitamin D uncorrelated with Covid severity in those of Eu... by tgb on Thursday January 21, 2021 @09:42AM (#25859277)
This is an interesting perspective I haven't considered. Isn't the exposure here the Vitamin deficiency? And so that could have chronic effects even though the infection is acute.