Allele frequency - compare groups --within

# Allele frequency comparisons with PLINK
# Comparisons between multiple groups using --within

# Clear workspace
rm(list = ls())
# Set working directory
setwd("d:/analysis/2020_GenomicsBootCamp_Demo/")
library(tidyverse)

# Read in the data - ADAPTmap
goats <- read_tsv("ADAPTmap_genotypeTOP_20160222_full.fam", col_names = F)

# select goats from target countries and apply country label
countryLabels <- goats %>%
  mutate(country = case_when(
    substr(X2,1,2) == "ET" ~ "Ethiopia",
    substr(X2,1,2) == "IT" ~ "Italy",
    substr(X2,1,2) == "PK" ~ "Pakistan"
  )) %>%
  drop_na() %>%
  select(X1, X2, country) %>%
  write_delim("countryComparisonSnpFreq.txt")


# Keep only these animals for a PLINK run and perform alle frequency analysis
system(str_c("plink --bfile ADAPTmap_genotypeTOP_20160222_full --chr-set 29 --autosome ",
              "--keep countryComparisonSnpFreq.txt --nonfounders ",
              "--freq --within countryComparisonSnpFreq.txt ",
              "--out countrySnpFrequencies"))