search

weizhou0 / qtl

GNU General Public License v3.0
19 stars 2 forks source link

Plink question #12

Open liuxiaZzz opened 1 month ago

liuxiaZzz commented 1 month ago

Hi, thanks for your excellent tool! I am wondering about the choice of plink. Should I use the same chr in step1 and step2&3 or just randomly choose in step1? Because I noticed that step1 used chr1 plink while step2 used chr2 plink in the tutorial. Thanks for your patience!

drneavin commented 2 weeks ago

I second this question. I'm a bit unclear on the differences are for the genotype data used in step 1 vs step 2. Any chance you could provide clarification @weizhou0 @weizhouUMICH ?

annacuomo commented 2 weeks ago

Hi both, sorry for the confusion. The genotype files used in the two different steps are different!

The plink file used in step 1 should be the same across the entire analysis, and uses a random subset of markers for the estimation of the variance ration when fitting the null model. As specified here you should "include 1000 random markers with MAC >= 20". We'd recommend choosing variants across all chromosomes and after pruning to have the most representative set possible.

The genotype file for steps 2&3 should contain the variants you want to test for that specific phenotype / gene, and the chromosome flag specifies the chromosome for the relevant gene + variants.

I hope this helps!