# Lab 6 for the University of Tulsa's CS-6643 Bioinformatics Course
# GWAS
# Professor: Dr. McKinney, Fall 2022
# Noah L. Schrick - 1492657

## Set Working Directory to file directory - RStudio approach
setwd(dirname(rstudioapi::getActiveDocumentContext()$path))

#### Part 0: PLINK
if (!require("BiocManager")) install.packages("BiocManager")
library(BiocManager)
if (!require("snpStats")) BiocManager::install("snpStats")
library(snpStats)

ex.data <- read.pedfile(file="extra.ped", snps="extra.map")
ex.data$fam
phenotype <- ex.data$fam$affected-1  # change pheno from 1/2 to 0/1
genotypes <- ex.data$genotypes       # encoded as AA/AB/BB
snp.ids <- as.character(ex.data$map$snp.names)
genotypes.df <- data.frame(as(genotypes, "character"))
colnames(genotypes.df) <- snp.ids
# observed contingency table for SNP rs630969
table(phenotype,genotypes.df$rs630969,
      dnn=c("phenotype","genotype")) # dnn dimension names of table