I was recently in an introductory seminar about Genome Wide Association Studies (GWAS). These studies aim to examine if any genetic markers are associate with a certain trait. It is not uncommon to have millions of separate tests. It is common practice for GWAS to conduct replications before publication.
However, due to the limited statistical power of many studies, meta-analysis have been conducted to find genetic markers which are genome wide significant. The speaker in my seminar noted that those Meta-analysis do not require replications.
I don't really understand this logic and I was hoping someone could explain this to me. In my understanding, GWAS meta-analysis do not only consider replicated/validated GWAS signals but rather all tested genetic markers in order to find new genome wide associated variants. Shouldn't it be necessary to replicate these new findings?