search

YuanTian1991 / ChAMP

19 stars 22 forks source link

champ.DMR different results which are not expected #25

Open lorbaer opened 2 years ago

lorbaer commented 2 years ago

Hi, I was running champ.DMR recently using the bumphunter method and I got some unexpected results. My Dataset has 4 Groups. running:

try1 <- champ.DMR(arraytype = "EPIC", beta=myCombat_all, pheno=myLoad$pd$Sample_Group)

try2 <- champ.DMR(arraytype = "EPIC", beta=myCombat_all, pheno=myLoad$pd$Sample_Group, compare.group=c("A","B")

try3 <- champ.DMR(arraytype = "EPIC", beta=myCombat_onlyA_and_B, pheno=myLoad$pd$Sample_Group[which(myLoad$pd$Sample_Group=="A" | myLoad$pd$Sample_Group=="B")])

try4 <- champ.DMR(arraytype = "EPIC", beta=myCombat_onlyA_and_B, pheno=myLoad$pd$Sample_Group[which(myLoad$pd$Sample_Group=="A" | myLoad$pd$Sample_Group=="B")]), compare.group=c("A","B")

myCombat_onlyA_and_B is the filtered version of myCombal_all only including the samples from group A and B.

try1 = 127 DMRs try2 = 134 DMRs try3 = 34 DMRs try4 = 37 DMRs

I expected try1 to be different from the rest and maybe I could also expect try2 to be different to 3 and 4, but I definitely expected try3 and try4 to be identical.

Did I missunderstood something?