Effect allele frequency in PRS calculation

[Mary]

Hi Sam and all,

As I found, it is recommended to use SNPs with effect allele frequency (EAF) > 0.01 for PRS calculation; however, for some traits, such as type 1 diabetes, some papers published in valid journals like this at Utility of Diabetes Type–Specific Genetic Risk Scores for the Classification of Diabetes Type Among Multiethnic Youth - PMC (nih.gov), used multiple SNPs with EFA < 0.01. So, considering SNPs with low EAF would be OK for some traits, right?

Another point is whether should we consider SNPs with similar EAF between two populations for PRS calculation? 

Thank you