--indep-pairwise best practice?

[Oveis Jamialahmadi]

Hi Christopher!

I am aware that this question has been asked probably several times; but I'm interested to know more details if possible:

Is there any particular way to assess the best parameters for --indep-pairwise? I'm aware that it depends on downstream purposes and 'kb' can be used (e.g. 500 kb 1 r2) for more stringnet purposes, but still the results would be different. As an example, I ran PLINK2 using these 3 different scenarios:

1. --indep-pairwise 500 kb 1 0.2
2. --indep-pairwise 50 5 0.2
3. --indep-pairwise 1000 80 0.2

Each of them generates a different set of markers, with some being unique (e.g. approach 1 cannot detect markers above 500 kb distance). My question is: is there any approach to optimize these parameters? Let's say we want to detect all possible independent markers with a specific r2. Is this even possible or if not, is there any greedy approach to find them?

Thanks in advance,

Best/Oveis  

[Christopher Chang]

https://groups.google.com/d/msg/plink2-users/2_XCfTVgk8I/4Sc5OUHaCAAJ

You can scan an entire chromosome at a time by setting the window size larger than any chromosome length (e.g. "--indep-pairwise 250000kb ..."), but this is rarely worth the computational cost.