Question on missing data management

[jbruxaux]

Hi!

I have a quick question on the different ways to handle missing data with RAiSD.
As far as I understand, the default behavior (0) is to discard entirely any SNP with any amount of missing data.
Option 1 imputes any missing data. How is it done? Based on the allele frequency at this position? Only that or other parameters (nearby SNPs...)?
And I'm not sure to really understand options 2 and 3, or rather, what's the difference between them?

Thanks for your help!

Jade