phewas function give inconsistent results

[YuxiaoRuoyao]

Hi,

I found that phewas function could give different number of associations for the same variant. Basically input less variants will find more associations and traits so I'm wondering why. Here is a small example:

batch = c("ieu-a","ieu-b","ukb-b","ebi-a")
m1 <- ieugwasr::phewas(variants = "rs12132412",batch=batch)
id_name <- c("rs12132412","rs3935032", "rs585406", "rs469882", "rs2211320", "rs630372", "rs10924372", "rs12239046", "rs6695572", "rs78511209", "rs12730935", "rs55852476", "rs75460349", "rs3768321", "rs1353595", "rs197422", "rs17417252", "rs11118625", "rs11577023", "rs34298354", "rs301798", "rs34761529", "rs4655802", "rs2376015", "rs797680", "rs7528419", "rs34065906", "rs59519769", "rs2325924", "rs4433388")
m2 <- ieugwasr::phewas(variants = id_name,batch=batch)
m1
m2 %>% filter(rsid == "rs12132412")

So m1 actually has 17 associations but m2 with "rs12132412" only has 11 associations. This issue is not accidental and if inputting more variants in one time, less associations will be found. I'm confused about that and do you recommend we input variants one by one in order to get all associations satisfying the pvalue cutoff requirements?

Thank you!

Best,
Ruoyao