Cosbin/normalization.R at main · cbil-vt/Cosbin · 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
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148

#################################################### (group mean)

input_data <- data3
nGroup <- 3

#### Cosbin normaliztaion (group mean)
start.time <- Sys.time()
cosbin_out <- cosbin(input_data)
end.time <- Sys.time()
time.taken <- end.time - start.time
time.taken


#### Total count normaliztaion (group mean)
totalcount_out <- totalcount(input_data)$data


#### DESeq2 (group mean)
cond <- factor(1:nGroup)
input_data_int <- round(input_data)
dds <- DESeqDataSetFromMatrix(input_data_int, DataFrame(cond), ~ cond)
dds <- estimateSizeFactors(dds)
sizeFactors(dds) <- sizeFactors(dds)/geometric.mean(sizeFactors(dds))
DESeq2_out <- input_data
for(i in 1:nGroup){
  DESeq2_out[, i] <- DESeq2_out[, i] / sizeFactors(dds)[i]
}


#### TMM/edgeR (group mean)
normFactors <- edgeR::calcNormFactors(input_data)
normFactors <- normFactors/geometric.mean(normFactors)
tmm_out <- input_data
for(i in 1:nGroup){
  tmm_out[, i] <- tmm_out[, i] / normFactors[i]
}


#### TCC (group mean)
tcc <- new("TCC", input_data, 1:nGroup)
tcc <-
  calcNormFactors(
    tcc,
    norm.method = "tmm",
    test.method = "edger",
    iteration = 1,
    FDR = 0.1,
    floorPDEG = 0.05
  )
tccEsts <- tcc$norm.factors
tccEsts <- tccEsts / geometric.mean(tccEsts)

tcc_out <- input_data
for(i in 1:nGroup){
  tcc_out[, i] <- tcc_out[, i] / tccEsts[i]
}

#################################################### (with replicates, then average to evaluate)

input_data2 <- data2
nGroup <- 3
nRep <- c(10, 10, 10)

#### DESeq2 (with replicates, then average to evaluate)
cond <- factor(rep(1:nGroup, nRep))
input_data2_int <- round(input_data2)
dds <- DESeqDataSetFromMatrix(input_data2_int, DataFrame(cond), ~ cond)
dds <- estimateSizeFactors(dds)
sizeFactors(dds) <- sizeFactors(dds)/geometric.mean(sizeFactors(dds))
DESeq2_temp <- input_data2
for(i in 1:sum(nRep)){
  DESeq2_temp[, i] <- DESeq2_temp[, i] / sizeFactors(dds)[i]
}

## get average for evaluation
DESeq2_out <- NULL
start <- 1
for(rep in nRep){
  DESeq2_out <- cbind(DESeq2_out, rowMeans(DESeq2_temp[, start : (start + rep - 1)]))
  start <- start + rep
}

#### edgeR normalization - TMM (with replicates, then average to evaluate)
normFactors <- edgeR::calcNormFactors(input_data2)
totCounts <- colSums(input_data2)
normFactors <- normFactors * totCounts
normFactors <- normFactors/geometric.mean(normFactors)
tmm_temp <- input_data2
for(i in 1:sum(nRep)){
  tmm_temp[, i] <- tmm_temp[, i] / normFactors[i]
}

## get average for evaluation
tmm_out <- NULL
start <- 1
for(rep in nRep){
  tmm_out <- cbind(tmm_out, rowMeans(tmm_temp[, start : (start + rep - 1)]))
  start <- start + rep
}

#### TCC (group mean, different test.method)
tcc <- new("TCC", input_data, 1:nGroup)
tcc <-
  TCC::calcNormFactors(
    tcc,
    norm.method = "tmm",
    test.method = "wad",
    iteration = 1,
    FDR = 0.1,
    floorPDEG = 0.05
  )
tccEsts <- tcc$norm.factors
totCounts <- colSums(input_data)
tccEsts <- tccEsts * totCounts
tccEsts <- tccEsts / geometric.mean(tccEsts)

tcc_out <- getNormalizedData(tcc)


#### TCC (with replicates, then average to evaluate)
tcc <- new("TCC", input_data2, rep(1:nGroup, each = 10))
tcc <-
  calcNormFactors(
    tcc,
    norm.method = "tmm",
    test.method = "edger",
    iteration = 1,
    FDR = 0.1,
    floorPDEG = 0.05
  )
tccEsts <- tcc$norm.factors
totCounts <- colSums(input_data2)
tccEsts <- tccEsts * totCounts
tccEsts <- tccEsts / geometric.mean(tccEsts)

tcc_temp <- input_data2
for(i in 1:sum(nRep)){
  tcc_temp[, i] <- tcc_temp[, i] / tccEsts[i]
}

## get average for evaluation
tcc_out <- NULL
start <- 1
for(rep in nRep){
  tcc_out <- cbind(tcc_out, rowMeans(tcc_temp[, start : (start + rep - 1)]))
  start <- start + rep
}