LD reference optimization according to GWAS ancestry from input

[AMCalejandro]

Is your feature request related to a problem? Please describe.
We want to be able to get the exact superpopulation, or even better, the subpopulation from the input GWAS, by correlating the input GWAS MAFs against the the LDref MAFs.

Describe the solution you'd like

• Get the GWAS MAFs. Split into a) Genotyped, b) Geenotyped + Imputed and perform the correlation
• Get the most correlated superpopulation / subpopulation, and use the subset to perform fine-mapping

Describe alternatives you've considered
If the inputted superpopulation does not really mach, rise a warning for now

Additional context
Read what is out there and maybe try to perform an undustment od the LD reference (ie deconvolution methods)

NOTE
Go to translate_population.R