Simulated data

## Simulation
#library
library(dplyr)
library(Matrix)
library(SingleCellExperiment);
library(tibble);
library(Seurat);
library(ROCR);
library(cluster);
library(mclust)
library(M3Drop)
library(magrittr)
library(rprintf)
source('../PCA_entropy/pca_entropy.R')
#droplet-based data


library(splatter)
library(SIMLR)
library(igraph)
library(M3Drop)
library(tibble)
library(rprintf)
library(scater)
library(rsvd)
source('Rogue.R')
source('pca_entropy.R')

n_gene=500 
n_iter = 20
ARI_c <- matrix(NA,ncol = 5,nrow = n_iter)
colnames(ARI_c) <- c('random','IE','SE','HVG','seed')
ARI_c <- as.data.frame(ARI_c)
for(i in 1:n_iter)
{
cat(sprintf(" Performing the %s iteration", i))
seed <- ceiling(runif(1,min = 1, max = 1000000))
ARI_c[i,5] <- seed
sim <- splatSimulate(group.prob=c(0.05,0.15,0.2,0.2,0.4),method = 'group',verbose = F,
                     nGenes=20000,batchCells=500,seed = seed )
group <- sim$Group
count <- counts(sim)               
data <- count/colSums(count)*10000
s1 <- sample(rownames(data),n_gene)
ent_res <- SE_fun(count,r=0.001)
int_res <- Get_entropy(data,5,0.1)
m <- nrow(int_res)
cat(sprintf(" the number of int genes is %s \n", m))

HVG_res <- BrenneckeGetVariableGenes(count,fdr = 1.2,suppress.plot=TRUE)
s2 <- as.character(int_res$Gene[1:n_gene])
s3 <- as.character(ent_res$Gene[1:n_gene])
s4 <- as.character(HVG_res$Gene[1:n_gene])
simlr1 <- SIMLR_Large_Scale(data[rownames(data)%in%s1,], c = 5)
simlr2 <- SIMLR_Large_Scale(data[rownames(data)%in%s2,], c = 5)
simlr3 <- SIMLR_Large_Scale(data[rownames(data)%in%s3,], c = 5)
simlr4 <- SIMLR_Large_Scale(data[rownames(data)%in%s4,], c = 5)
ARI_c[i,1] <- compare(simlr1$y$cluster,group,method = 'adjusted.rand')
ARI_c[i,2] <- compare(simlr2$y$cluster,group,method = 'adjusted.rand')
ARI_c[i,3] <- compare(simlr3$y$cluster,group,method = 'adjusted.rand')
ARI_c[i,4] <- compare(simlr4$y$cluster,group,method = 'adjusted.rand')
}