####
#input file: two-column; raw genotype data; first row locus names; first column individual id
####
#### define number of simulated samples and iterations
no.sim.samples=15
n.its<-1000
#########

datasim<-read.table("eg_input.txt", header=T, sep="\t")
reboot=datasim
no.samples=dim(datasim)[1]
noloci=(dim(datasim)[2]-1)/2

population=datasim[,-c(1)] 

L=ncol(population)  
locus_positions=(2*(unique(round((1:(L-2))/2)))+1)   
lnames=colnames(population)                          
OUT=NULL        

for (x in locus_positions) {                       
  alleles=c(population[,x],population[,x+1])        
  alleles2=as.data.frame(table(alleles))            
        
  alleles3=alleles2[which(alleles2[,1]!=0),]          
  alleles4=cbind(alleles3,alleles3[,2]/sum(alleles3[,2])) 
  output=cbind(x,lnames[x],alleles4)                        
  
  OUT <<- rbind(OUT,output)
  
}


colnames(OUT) <- c("Number","Locus","allele","count","frequency") #add column headers

allalleles=dim(OUT)[1]

prop=data.frame(matrix(0,2,no.sim.samples))
prop[1,]=c(1:no.sim.samples)
collect<-data.frame(matrix(0,n.its,no.sim.samples))
names(collect)<-1:no.sim.samples
names(prop)<-1:no.sim.samples

tempX=as.data.frame(matrix(1,no.sim.samples,noloci*2))
names(tempX)=names(population)


for (d in 1:no.sim.samples){
print(d)
for(i in 1:n.its){

for(W in 1:noloci){

tempX[1:d,(W+(W-1))]=as.character(sample(OUT[which(OUT$Number==(W+(W-1))),3],d,repl=TRUE,prob=OUT[which(OUT$Number==(W+(W-1))),5]))
tempX[1:d,(W+(W-1)+1)]=as.character(sample(OUT[which(OUT$Number==(W+(W-1))),3],d,repl=TRUE,prob=OUT[which(OUT$Number==(W+(W-1))),5]))
}


loc1_5<-length(sort(unique(c(unique(tempX[,1][1:d]),unique(tempX[,2][1:d])))))
loc2_5<-length(sort(unique(c(unique(tempX[,3][1:d]),unique(tempX[,4][1:d])))))
loc3_5<-length(sort(unique(c(unique(tempX[,5][1:d]),unique(tempX[,6][1:d])))))
loc4_5<-length(sort(unique(c(unique(tempX[,7][1:d]),unique(tempX[,8][1:d])))))
loc5_5<-length(sort(unique(c(unique(tempX[,9][1:d]),unique(tempX[,10][1:d])))))
loc6_5<-length(sort(unique(c(unique(tempX[,11][1:d]),unique(tempX[,12][1:d])))))
loc7_5<-length(sort(unique(c(unique(tempX[,13][1:d]),unique(tempX[,14][1:d])))))
loc8_5<-length(sort(unique(c(unique(tempX[,15][1:d]),unique(tempX[,16][1:d])))))
loc9_5<-length(sort(unique(c(unique(tempX[,17][1:d]),unique(tempX[,18][1:d])))))
loc10_5<-length(sort(unique(c(unique(tempX[,19][1:d]),unique(tempX[,20][1:d])))))
loc11_5<-length(sort(unique(c(unique(tempX[,21][1:d]),unique(tempX[,22][1:d])))))
loc12_5<-length(sort(unique(c(unique(tempX[,23][1:d]),unique(tempX[,24][1:d])))))
loc13_5<-length(sort(unique(c(unique(tempX[,25][1:d]),unique(tempX[,26][1:d])))))
loc14_5<-length(sort(unique(c(unique(tempX[,27][1:d]),unique(tempX[,28][1:d])))))
loc15_5<-length(sort(unique(c(unique(tempX[,29][1:d]),unique(tempX[,30][1:d])))))
loc16_5<-length(sort(unique(c(unique(tempX[,31][1:d]),unique(tempX[,32][1:d])))))
loc17_5<-length(sort(unique(c(unique(tempX[,33][1:d]),unique(tempX[,34][1:d])))))
loc18_5<-length(sort(unique(c(unique(tempX[,35][1:d]),unique(tempX[,36][1:d])))))
loc19_5<-length(sort(unique(c(unique(tempX[,37][1:d]),unique(tempX[,38][1:d])))))
loc20_5<-length(sort(unique(c(unique(tempX[,39][1:d]),unique(tempX[,40][1:d])))))

no.alleles_5<-loc1_5+loc2_5+loc3_5+loc4_5+loc5_5+loc6_5+loc7_5+loc8_5+loc9_5+loc10_5+loc11_5+
+loc12_5+loc13_5+loc14_5+loc15_5+loc16_5+loc17_5+loc18_5+loc19_5+loc20_5


prop_samp<-no.alleles_5/allalleles

collect[i,d]<-prop_samp
}
}


for(z in 1:no.sim.samples){

prop[2,z]=length(collect[,z][collect[,z]==1])/n.its
}

results_allloci=(rbind(prop,collect))

write.table(results_allloci,"results_allloci_allelefrq.txt")

plot(as.numeric(prop[2,])~as.numeric((prop[1,])),xlab="no. of individuals subsampled",ylab=("proportion of data sets capturing total no. of alleles"))

abline(h=0.05,lty=2)



###############
# remove freq <0.02
##########


population=datasim[,-c(1)] 

L=ncol(population)  
locus_positions=(2*(unique(round((1:(L-2))/2)))+1)   
lnames=colnames(population)                           
OUT=NULL        

for (x in locus_positions) {                      
  alleles=c(population[,x],population[,x+1])        
  alleles2=as.data.frame(table(alleles))             

  alleles3=alleles2[which(alleles2[,1]!=0),]          
  alleles4=cbind(alleles3,alleles3[,2]/sum(alleles3[,2])) 
  output=cbind(x,lnames[x],alleles4)                        

  
  OUT <<- rbind(OUT,output)
  
}


colnames(OUT) <- c("Number","Locus","allele","count","frequency")


OUT2=OUT[which(OUT$frequency<0.02),] 

nolociX=dim(OUT2)[1]


for (i in 1:nolociX) {   

population=population[which(population[,OUT2$Number[i]]!= OUT2$allele[i]),]
population=population[which(population[,OUT2$Number[i]+1]!= OUT2$allele[i]),]

}


datasimbb=population



OUT3=NULL  
for (x in locus_positions) {                      
  alleles=c(datasimbb[,x],datasimbb[,x+1])        
  alleles2=as.data.frame(table(alleles))             
            
  alleles3=alleles2[which(alleles2[,1]!=0),]          
  alleles4=cbind(alleles3,alleles3[,2]/sum(alleles3[,2])) 
  output=cbind(x,lnames[x],alleles4)

  
  OUT3 <<- rbind(OUT3,output)
  
}


colnames(OUT3) <- c("Number","Locus","allele","count","frequency") 


allalleles=dim(OUT3)[1]

prop=data.frame(matrix(0,2,no.sim.samples))
prop[1,]=c(1:no.sim.samples)
collect<-data.frame(matrix(0,n.its,no.sim.samples))
names(collect)<-1:no.sim.samples
names(prop)<-1:no.sim.samples

tempX=as.data.frame(matrix(1,no.sim.samples,noloci*2))
names(tempX)=names(population)


for (d in 1:no.sim.samples){
print(d)
for(i in 1:n.its){

for(W in 1:noloci){

tempX[1:d,(W+(W-1))]=as.character(sample(OUT3[which(OUT3$Number==(W+(W-1))),3],d,repl=TRUE,prob=OUT3[which(OUT3$Number==(W+(W-1))),5]))
tempX[1:d,(W+(W-1)+1)]=as.character(sample(OUT3[which(OUT3$Number==(W+(W-1))),3],d,repl=TRUE,prob=OUT3[which(OUT3$Number==(W+(W-1))),5]))
}


loc1_5<-length(sort(unique(c(unique(tempX[,1][1:d]),unique(tempX[,2][1:d])))))
loc2_5<-length(sort(unique(c(unique(tempX[,3][1:d]),unique(tempX[,4][1:d])))))
loc3_5<-length(sort(unique(c(unique(tempX[,5][1:d]),unique(tempX[,6][1:d])))))
loc4_5<-length(sort(unique(c(unique(tempX[,7][1:d]),unique(tempX[,8][1:d])))))
loc5_5<-length(sort(unique(c(unique(tempX[,9][1:d]),unique(tempX[,10][1:d])))))
loc6_5<-length(sort(unique(c(unique(tempX[,11][1:d]),unique(tempX[,12][1:d])))))
loc7_5<-length(sort(unique(c(unique(tempX[,13][1:d]),unique(tempX[,14][1:d])))))
loc8_5<-length(sort(unique(c(unique(tempX[,15][1:d]),unique(tempX[,16][1:d])))))
loc9_5<-length(sort(unique(c(unique(tempX[,17][1:d]),unique(tempX[,18][1:d])))))
loc10_5<-length(sort(unique(c(unique(tempX[,19][1:d]),unique(tempX[,20][1:d])))))
loc11_5<-length(sort(unique(c(unique(tempX[,21][1:d]),unique(tempX[,22][1:d])))))
loc12_5<-length(sort(unique(c(unique(tempX[,23][1:d]),unique(tempX[,24][1:d])))))
loc13_5<-length(sort(unique(c(unique(tempX[,25][1:d]),unique(tempX[,26][1:d])))))
loc14_5<-length(sort(unique(c(unique(tempX[,27][1:d]),unique(tempX[,28][1:d])))))
loc15_5<-length(sort(unique(c(unique(tempX[,29][1:d]),unique(tempX[,30][1:d])))))
loc16_5<-length(sort(unique(c(unique(tempX[,31][1:d]),unique(tempX[,32][1:d])))))
loc17_5<-length(sort(unique(c(unique(tempX[,33][1:d]),unique(tempX[,34][1:d])))))
loc18_5<-length(sort(unique(c(unique(tempX[,35][1:d]),unique(tempX[,36][1:d])))))
loc19_5<-length(sort(unique(c(unique(tempX[,37][1:d]),unique(tempX[,38][1:d])))))
loc20_5<-length(sort(unique(c(unique(tempX[,39][1:d]),unique(tempX[,40][1:d])))))

no.alleles_5<-loc1_5+loc2_5+loc3_5+loc4_5+loc5_5+loc6_5+loc7_5+loc8_5+loc9_5+loc10_5+loc11_5+
+loc12_5+loc13_5+loc14_5+loc15_5+loc16_5+loc17_5+loc18_5+loc19_5+loc20_5


prop_samp<-no.alleles_5/allalleles

collect[i,d]<-prop_samp
}
}


for(z in 1:no.sim.samples){

prop[2,z]=length(collect[,z][collect[,z]==1])/n.its
}

results_allloci=(rbind(prop,collect))

write.table(results_allloci,"results_nolowfreq_alleles.txt")

plot(as.numeric(prop[2,])~as.numeric((prop[1,])),xlab="no. of individuals subsampled",ylab=("proportion of data sets capturing total no. of alleles"))

abline(h=0.05,lty=2)




###################################
##   leave samples with private alleles in
#################


datasim<-reboot
priv_all=read.table("sampwprivall.txt", header=T, sep="\t")
priv_all=as.data.frame(priv_all[,1])


###############
# remove freq <0.02
##########

population=datasim
for(B in 1:(length(priv_all[,1]))){
population=population[which(population[,1]!= as.character(priv_all[B,1])),]         
}
dim(population)

excluded=datasim
toexcl=as.data.frame(population[,1])

for(B in 1:(length(toexcl[,1]))){
excluded=excluded[which(excluded[,1]!= as.character(toexcl[B,1])),]         
}

########
# check for errors
########

dim(datasim)[1]-dim(population)[1]-dim(excluded)[1]

population=population[,-c(1)] 

L=ncol(population)   
locus_positions=(2*(unique(round((1:(L-2))/2)))+1)   
lnames=colnames(population)                          
OUT=NULL        

for (x in locus_positions) {                       
  alleles=c(population[,x],population[,x+1])        
  alleles2=as.data.frame(table(alleles))            
        
  alleles3=alleles2[which(alleles2[,1]!=0),]          
  alleles4=cbind(alleles3,alleles3[,2]/sum(alleles3[,2])) 
  output=cbind(x,lnames[x],alleles4)                        
  
  OUT <<- rbind(OUT,output)
  
}


colnames(OUT) <- c("Number","Locus","allele","count","frequency") 

OUT2=OUT[which(OUT$frequency<0.02),] 

noXloci=dim(OUT2)[1]


for (i in 1:noXloci) {   

population=population[which(population[,OUT2$Number[i]]!= OUT2$allele[i]),]
population=population[which(population[,OUT2$Number[i]+1]!= OUT2$allele[i]),]

}

dim(population)

population=rbind(population,excluded[,-c(1)] )


datasimbb=population

dim(population)



OUT3=NULL  
for (x in locus_positions) {                       
  alleles=c(datasimbb[,x],datasimbb[,x+1])        
  alleles2=as.data.frame(table(alleles))             
         
  alleles3=alleles2[which(alleles2[,1]!=0),]          
  alleles4=cbind(alleles3,alleles3[,2]/sum(alleles3[,2])) 
  output=cbind(x,lnames[x],alleles4)                        
  
  OUT3 <<- rbind(OUT3,output)
  
}


colnames(OUT3) <- c("Number","Locus","allele","count","frequency") 


allalleles=dim(OUT3)[1]

prop=data.frame(matrix(0,2,no.sim.samples))
prop[1,]=c(1:no.sim.samples)
collect<-data.frame(matrix(0,n.its,no.sim.samples))
names(collect)<-1:no.sim.samples
names(prop)<-1:no.sim.samples

tempX=as.data.frame(matrix(1,no.sim.samples,noloci*2))
names(tempX)=names(population)


for (d in 1:no.sim.samples){
print(d)
for(i in 1:n.its){

for(W in 1:noloci){

tempX[1:d,(W+(W-1))]=as.character(sample(OUT3[which(OUT3$Number==(W+(W-1))),3],d,repl=TRUE,prob=OUT3[which(OUT3$Number==(W+(W-1))),5]))
tempX[1:d,(W+(W-1)+1)]=as.character(sample(OUT3[which(OUT3$Number==(W+(W-1))),3],d,repl=TRUE,prob=OUT3[which(OUT3$Number==(W+(W-1))),5]))
}


loc1_5<-length(sort(unique(c(unique(tempX[,1][1:d]),unique(tempX[,2][1:d])))))
loc2_5<-length(sort(unique(c(unique(tempX[,3][1:d]),unique(tempX[,4][1:d])))))
loc3_5<-length(sort(unique(c(unique(tempX[,5][1:d]),unique(tempX[,6][1:d])))))
loc4_5<-length(sort(unique(c(unique(tempX[,7][1:d]),unique(tempX[,8][1:d])))))
loc5_5<-length(sort(unique(c(unique(tempX[,9][1:d]),unique(tempX[,10][1:d])))))
loc6_5<-length(sort(unique(c(unique(tempX[,11][1:d]),unique(tempX[,12][1:d])))))
loc7_5<-length(sort(unique(c(unique(tempX[,13][1:d]),unique(tempX[,14][1:d])))))
loc8_5<-length(sort(unique(c(unique(tempX[,15][1:d]),unique(tempX[,16][1:d])))))
loc9_5<-length(sort(unique(c(unique(tempX[,17][1:d]),unique(tempX[,18][1:d])))))
loc10_5<-length(sort(unique(c(unique(tempX[,19][1:d]),unique(tempX[,20][1:d])))))
loc11_5<-length(sort(unique(c(unique(tempX[,21][1:d]),unique(tempX[,22][1:d])))))
loc12_5<-length(sort(unique(c(unique(tempX[,23][1:d]),unique(tempX[,24][1:d])))))
loc13_5<-length(sort(unique(c(unique(tempX[,25][1:d]),unique(tempX[,26][1:d])))))
loc14_5<-length(sort(unique(c(unique(tempX[,27][1:d]),unique(tempX[,28][1:d])))))
loc15_5<-length(sort(unique(c(unique(tempX[,29][1:d]),unique(tempX[,30][1:d])))))
loc16_5<-length(sort(unique(c(unique(tempX[,31][1:d]),unique(tempX[,32][1:d])))))
loc17_5<-length(sort(unique(c(unique(tempX[,33][1:d]),unique(tempX[,34][1:d])))))
loc18_5<-length(sort(unique(c(unique(tempX[,35][1:d]),unique(tempX[,36][1:d])))))
loc19_5<-length(sort(unique(c(unique(tempX[,37][1:d]),unique(tempX[,38][1:d])))))
loc20_5<-length(sort(unique(c(unique(tempX[,39][1:d]),unique(tempX[,40][1:d])))))

no.alleles_5<-loc1_5+loc2_5+loc3_5+loc4_5+loc5_5+loc6_5+loc7_5+loc8_5+loc9_5+loc10_5+loc11_5+
+loc12_5+loc13_5+loc14_5+loc15_5+loc16_5+loc17_5+loc18_5+loc19_5+loc20_5


prop_samp<-no.alleles_5/allalleles

collect[i,d]<-prop_samp
}
}


for(z in 1:no.sim.samples){

prop[2,z]=length(collect[,z][collect[,z]==1])/n.its
}

results_allloci=(rbind(prop,collect))

write.table(results_allloci,"results_nolowfreqnoprivate_alleles.txt")

plot(as.numeric(prop[2,])~as.numeric((prop[1,])),xlab="no. of individuals subsampled",ylab=("proportion of data sets capturing total no. of alleles"))

abline(h=0.05,lty=2)