tsImpute/example.R at main · ZhengWeihuaYNU/tsImpute · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
#Seurat, factoextra and umap are prerequisite packages for tsImpute, make sure that you have installed
#these packages before you install tsImpute.
#install.packages('Seurat')
#install.packages('factoextra')
#install.packages('umap')
#install.packages('flexclust')
install.packages("your path of tsImpute_0.1.1.tar.gz", repos = NULL, type = "source")
library(Seurat)
library(factoextra)
library(umap)
library(flexclust) #This package is not necessary for tsImpute, we use it to calculate adjusted Rand Index
library(parallel)
library(tsImpute)
#setwd('Your path of the data')
#read data and corresponding labels:
dat<- readRDS('darmanis.rds')
cell_types<- readRDS('cell_types.rds')

#Users can also try tsimpute with Ting data:
#dat<- readRDS('ting.rds')
#cell_types<- readRDS('cell_types_ting.rds')

#run tsimpute:
output<- tsimpute(dat, seed= 1)


#use Seurat package to cluster the cells and calculate adjusted Rand index (ARI):
x.seurat <- CreateSeuratObject(output)
x.seurat <- NormalizeData(x.seurat)
x.seurat <- ScaleData(x.seurat)
x.seurat <- FindVariableFeatures(x.seurat, verbose = FALSE)
x.seurat <- RunPCA(x.seurat, features = VariableFeatures(object = x.seurat), seed.use = 1)
x.seurat <- JackStraw(x.seurat, num.replicate = 100)
x.seurat <- ScoreJackStraw(x.seurat, dims = 1:20)
x.seurat <- FindNeighbors(x.seurat, dims = 1:10) #Use the first 10 components for clustering
x.seurat <- FindClusters(x.seurat, random.seed = 1)
final_cluster<- x.seurat$seurat_clusters
ari<- randIndex(cell_types, final_cluster) #ARI
print(ari) #0.800

#Compare it to the clustering results of raw data:
x.seurat <- CreateSeuratObject(dat)
x.seurat <- NormalizeData(x.seurat)
x.seurat <- ScaleData(x.seurat)
x.seurat <- FindVariableFeatures(x.seurat, verbose = FALSE)
x.seurat <- RunPCA(x.seurat, features = VariableFeatures(object = x.seurat), seed.use = 1)
x.seurat <- JackStraw(x.seurat, num.replicate = 100)
x.seurat <- ScoreJackStraw(x.seurat, dims = 1:20)
x.seurat <- FindNeighbors(x.seurat, dims = 1:10) #Use the first 10 components for clustering
x.seurat <- FindClusters(x.seurat, random.seed = 1)
raw_cluster<- x.seurat$seurat_clusters
raw_ari<- randIndex(cell_types, raw_cluster) #ARI
print(raw_ari) #0.663