##### File S1. Script to generate example simulated datasets and impute missing data. #####

require("devtools")
library(devtools)
install_github(user="behuang", repo="mpMap")
library(mpMap)

NumberMarkers <- 201
NumberProgeny <- 200
MissingProgenyProb <- .6
MissingParentProb <- .36
ChromosomeLength <- 100
seed <- 1

set.seed(seed)
map <- sim.map(len=ChromosomeLength, n.mar=NumberMarkers,  eq.spacing=T, include.x=F)
## Generate 8-way pedigree
ped <- sim.mpped(nfounders=8, nfunnels=1, nperfam=NumberProgeny, nssdgen=6, nseeds=1)

## Simulate full dataset
mpobj <- sim.mpcross(map=map, pedigree=ped, seed=seed)
FullSimData <- rbind(mpobj$founders, mpobj$finals)
fullmp <- mpobj

## Simulate which progeny genotypes will be missing
missmat <- matrix(rbinom(prod(dim(mpobj$finals)),1, p=MissingProgenyProb), nrow=nrow(mpobj$finals), ncol=ncol(mpobj$finals))
mpobj$finals[missmat==1] <- NA

## Simulate which parent genotypes will be missing
missfou <- matrix(rbinom(prod(dim(mpobj$founders)), 1, p=MissingParentProb), nrow=nrow(mpobj$founders), ncol=ncol(mpobj$founders))
mpobj$founders[missfou==1] <- NA

MissSimData <- rbind(mpobj$founders, mpobj$finals)

imputedmp <- mpimpute(mpobj, what="both")

## How many missing in parents? progeny? ## check this - should have better performance. 
sum(is.na(imputedmp$founders))

## How many missing in progeny?
sum(is.na(imputedmp$finals))

## Proportion incorrectly imputed in parents? progeny?
sum(imputedmp$founders[which(is.na(mpobj$founders))]!=fullmp$founders[which(is.na(mpobj$founders))])/sum(is.na(mpobj$founders))
sum(imputedmp$finals[which(is.na(mpobj$finals))]!=fullmp$finals[which(is.na(mpobj$finals))], na.rm=T)/sum(is.na(mpobj$finals))