What are the default priors to compute PPA (posterior probability of association?

[Shuwei]

Dear Yongtao,

If I use a 2-SNP model, the summary file will give me a matrix of BF
and a corresponding matrix of PPA. I know that PPA=PO/(1+PO), and
PO=BF*pi/(1-pi)

But I don't know what is the default setting for pi in BIMBAM?
It is said p(l) proportional to 0.5^l, so does it mean pi=0.5 for 1-
SNP model and pi=0.25 for 2-SNP model?

But the result told me it is not right.
Seems that pi=3.5e-5 for each SNP in 1-SNP model and pi=8.23e-7 for
each SNP in 2-SNP model

There are 61 SNPs in a region

I searched and the previous related answer was below, which I don't
quite understand.
> To get the posterior probability, do you choose the default prior at
> 0.5, the multiply the BF by this?

The posterior prob is computed from:
posterior proportional to prior times BF.

Incidentally there is a bug in the output of the multi-SNP BFs in the
summary file, which
will be fixed in the next release (the 1-SNP BF is multiplied by 0.5,
the 2-SNP BF
is wrongly multiplied by 0.25...)