how to calculate a MAF from multiple cohorts

[Royce Clifford]

This is a wonderful conversation group, very helpful.

I have sumstats from 4 cohorts, with a combined odds ratio for the meta-analysis but separate minor allele frequencies from each of the 4 cohorts. 

1. How shall I calculate the overall MAF? I'm told weighted samples would not be good... Or should I use the 1000 genomes values for a MAF?

Here is the header of the sumstats given for 2 of the cohorts, and I have the case/control numbers for each cohort:

RSID     chr     BP     EA     OA     ICE_freq     ICE_info     UK_freq     UK_info ...      OR     P

2. What should I use for Neff? The sum of individual Neffs, as seen in example 1  for the anxiety (ANX) example?

Thanks for any help you can give me.

Royce