PTSD GWAS: Another fast and loose with the term "Replication"

This study, a GWAS for PTSD (or symptoms related to PTSD, might be more accurate), like most found a few loci with significant p values. They found 8 in a sample of 120,000 "European Americans" and zero in a sample of 20,000 African Americans. I'd say this is what you might expect for the false positive rates of such studies. They go on to claim that it was "replicated in the UK Biobank." I'll make a quick point after the fold:
This is a table of the p-values of significance in their study and the p-values of the claimed replication in the UK Biobank:

As one can see, there is no p value of significance (5 X 10 -8 or lower) amongst the UK Biobank "replications." In fact, none is within 10,000 fold of a significant p value. Thus none of these were actually replicated. They are making hay of the fact that these p-values are in the same direction. Call it what you will, but that isn't replication.