-
Notifications
You must be signed in to change notification settings - Fork 3
/
diffASE.R
243 lines (221 loc) · 12.5 KB
/
diffASE.R
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
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
require("data.table")
require("metap")
require("limma")
require("biomaRt")
#mapping between gene names
ensembl = useEnsembl(biomart="ensembl", dataset="hsapiens_gene_ensembl")
geneMap = getBM(attributes=c("ensembl_gene_id", "hgnc_symbol", "entrezgene_id"), mart = ensembl)
#purity file with sample names in first column and purity in second
allPurity = fread("purity.csv")
#for differential expression, cpm matrices for tumor and normal where columns are samples and rows are genes, we assume they are ensembl transcipt ids (as produced by TCGA)
cpmTumor = read.table("cpmTumor.csv", row.names=1)
cpmNormal = read.table("cpmNormal.csv", row.names=1)
#ASE info has the first 7 columns as follows:
#Gene name, sample name, variant location, normal major read count, normal total read count, tumor major read count, tumor total read count
computeASEbaseline = function(ASEinfo){
#if multiple samples are being tested together
samples = unique(ASEinfo$V2)
#trim sites with no tumor reads
ASEinfo = ASEinfo[ASEinfo$V7!=0,]
#compute pvalues per heterozygous site
ASEraw_pval = c()
for(x in 1:dim(ASEinfo)[1]){
downsampledCount = min(ASEinfo$V7[x],100)
cancer_reads_raw = downsampledCount
pNormal = ASEinfo$V4[x]/ASEinfo$V5[x]
pCancer_raw = ASEinfo$V6[x]/ASEinfo$V7[x]
ASEraw_pval = c(ASEraw_pval, 2*min(pnorm(pCancer_raw*cancer_reads_raw, pNormal*cancer_reads_raw, sqrt(cancer_reads_raw*pNormal*(1-pNormal))),pnorm(pCancer_raw*cancer_reads_raw, pNormal*cancer_reads_raw, sqrt(cancer_reads_raw*pNormal*(1-pNormal)), lower.tail = FALSE)))
}
#Combine p-values, downsampling as needed
SampSigdown10 = c()
allTestGenes = c()
allSamples = c()
for(sample in samples){
testGenes = unique(ASEinfo$V1[ASEinfo$V2==sample])
allTestGenes = c(allTestGenes, testGenes)
allSamples = c(allSamples, rep(sample, length(testGenes)))
for(gene in testGenes){
if(length(which(ASEinfo$V1==gene & ASEinfo$V2==sample))==0){SampSigdown10=c(SampSigdown10,NA)}
else if(length(which(ASEinfo$V1==gene & ASEinfo$V2==sample))==1){SampSigdown10=c(SampSigdown10,min(ASEraw_pval[ASEinfo$V1==gene & ASEinfo$V2==sample],1))}
else if(length(which(ASEinfo$V1==gene & ASEinfo$V2==sample))>10){
SampSigdown10 = c(SampSigdown10,median(sapply(1:10, function(q) sumlog(sample(sapply(ASEraw_pval[ASEinfo$V1==gene & ASEinfo$V2==sample], function(x) min(x,1)), 10))[['p']])))
}
else{SampSigdown10 = c(SampSigdown10,sumlog(sapply(ASEraw_pval[ASEinfo$V1==gene & ASEinfo$V2==sample], function(x) min(x,1)))[['p']])}
}
}
#Compute weighted ASE
ASEsW100 = c()
for(sample in samples){
testGenes = unique(ASEinfo$V1[ASEinfo$V2==sample])
ASEsW100 = c(ASEsW100, sapply(testGenes, function(gene) weighted.median(abs((ASEinfo$V6/ASEinfo$V7-ASEinfo$V4/ASEinfo$V5)[ASEinfo$V1==gene & ASEinfo$V2==sample]), sapply(ASEinfo$V7[ASEinfo$V1==gene & ASEinfo$V2==sample], function(x) min(x,100)))))
}
data.frame(sample=allSamples, gene=allTestGenes, pval=SampSigdown10, ASE=ASEsW100)
}
computeASEpurity = function(ASEinfo){
#if multiple samples are being tested together
samples = unique(ASEinfo$V2)
#trim sites with no tumor reads
ASEinfo = ASEinfo[ASEinfo$V7!=0,]
#get purity scores for each ASE
asePurity = allPurity$V2[match(ASEinfo$V2,allPurity$V1)]
#compute purity corrected Rc
Rc = (ASEinfo$V6/ASEinfo$V7-(1-asePurity)*ASEinfo$V4/ASEinfo$V5)/asePurity
Rc = sapply(Rc, function(x) max(0,min(x,1)))
#compute pvalues per heterozygous site
ASErawPurity_pval = c()
for(x in 1:dim(ASEinfo)[1]){
downsampledCount = min(ASEinfo$V7[x],100)
cancer_reads_rawPurity = asePurity[x]*downsampledCount
pNormal = ASEinfo$V4[x]/ASEinfo$V5[x]
pCancer_rawPurity = Rc[x]
ASErawPurity_pval = c(ASErawPurity_pval, 2*min(pnorm(pCancer_rawPurity*cancer_reads_rawPurity, pNormal*cancer_reads_rawPurity, sqrt(cancer_reads_rawPurity*pNormal*(1-pNormal))),pnorm(pCancer_rawPurity*cancer_reads_rawPurity, pNormal*cancer_reads_rawPurity, sqrt(cancer_reads_rawPurity*pNormal*(1-pNormal)), lower.tail = FALSE)))
}
#Combine p-values, downsampling as needed
SampSigdown10 = c()
allTestGenes = c()
allSamples = c()
for(sample in samples){
testGenes = unique(ASEinfo$V1[ASEinfo$V2==sample])
allTestGenes = c(allTestGenes, testGenes)
allSamples = c(allSamples, rep(sample, length(testGenes)))
for(gene in testGenes){
if(length(which(ASEinfo$V1==gene & ASEinfo$V2==sample))==0){SampSigdown10=c(SampSigdown10,NA)}
else if(length(which(ASEinfo$V1==gene & ASEinfo$V2==sample))==1){SampSigdown10=c(SampSigdown10,min(ASErawPurity_pval[ASEinfo$V1==gene & ASEinfo$V2==sample],1))}
else if(length(which(ASEinfo$V1==gene & ASEinfo$V2==sample))>10){
SampSigdown10 = c(SampSigdown10,median(sapply(1:10, function(q) sumlog(sample(sapply(ASErawPurity_pval[ASEinfo$V1==gene & ASEinfo$V2==sample], function(x) min(x,1)), 10))[['p']])))
}
else{SampSigdown10 = c(SampSigdown10,sumlog(sapply(ASErawPurity_pval[ASEinfo$V1==gene & ASEinfo$V2==sample], function(x) min(x,1)))[['p']])}
}
}
#Compute weighted ASE
ASEsW100 = c()
for(sample in samples){
testGenes = unique(ASEinfo$V1[ASEinfo$V2==sample])
ASEsW100 = c(ASEsW100, sapply(testGenes, function(gene) weighted.median(abs((Rc-ASEinfo$V4/ASEinfo$V5)[ASEinfo$V1==gene & ASEinfo$V2==sample]), sapply(ASEinfo$V7[ASEinfo$V1==gene & ASEinfo$V2==sample], function(x) min(x,100)))))
}
data.frame(sample=allSamples, gene=allTestGenes, pval=SampSigdown10, ASE=ASEsW100)
}
computeASEexp = function(ASEinfo){
#if multiple samples are being tested together
samples = unique(ASEinfo$V2)
#trim sites with no tumor reads
ASEinfo = ASEinfo[ASEinfo$V7!=0,]
#get purity scores for each site
asePurity = allPurity$V2[match(ASEinfo$V2,allPurity$V1)]
#get expression levels for each site
aseExp = c()
for(i in 1:length(ASEinfo$V1)){
aseExp = rbind(aseExp, c(sum(cpmNormal[rownames(cpmNormal) %in% geneMap$ensembl_gene_id[match(ASEinfo$V1[i],geneMap$entrezgene)],colnames(cpmNormal)==ASEinfo$V2[i]]),sum(cpmTumor[rownames(cpmTumor) %in% geneMap$ensembl_gene_id[match(ASEinfo$V1[i],geneMap$entrezgene)],colnames(cpmTumor)==ASEinfo$V2[i]])))
}
#compute fraction of transcripts
ft = (aseExp[,2] - (1 - asePurity)*aseExp[,1])/(aseExp[,2])
ft = sapply(ft, function(x) max(0,min(x,1)))
#compute purity corrected Rc
Rc = (ASEinfo$V6/ASEinfo$V7-(1-ft)*ASEinfo$V4/ASEinfo$V5)/ft
Rc = sapply(Rc, function(x) max(0,min(x,1)))
#compute pvalues per heterozygous site
ASErawExp_pval = c()
for(x in 1:dim(ASEinfo)[1]){
downsampledCount = min(ASEinfo$V7[x],100)
cancer_reads_rawExp = ft[x]*downsampledCount
pNormal = ASEinfo$V4[x]/ASEinfo$V5[x]
pCancer_rawExp = Rc[x]
ASErawExp_pval = c(ASErawExp_pval, 2*min(pnorm(pCancer_rawExp*cancer_reads_rawExp, pNormal*cancer_reads_rawExp, sqrt(cancer_reads_rawExp*pNormal*(1-pNormal))),pnorm(pCancer_rawExp*cancer_reads_rawExp, pNormal*cancer_reads_rawExp, sqrt(cancer_reads_rawExp*pNormal*(1-pNormal)), lower.tail = FALSE)))
}
#Combine p-values, downsampling as needed
SampSigdown10 = c()
allTestGenes = c()
allSamples = c()
for(sample in samples){
testGenes = unique(ASEinfo$V1[ASEinfo$V2==sample])
allTestGenes = c(allTestGenes, testGenes)
allSamples = c(allSamples, rep(sample, length(testGenes)))
for(gene in testGenes){
if(length(which(ASEinfo$V1==gene & ASEinfo$V2==sample & !is.na(ASErawExp_pval)))==0){SampSigdown10=c(SampSigdown10,NA)}
else if(length(which(ASEinfo$V1==gene & ASEinfo$V2==sample & !is.na(ASErawExp_pval)))==1){SampSigdown10=c(SampSigdown10,min(ASErawExp_pval[ASEinfo$V1==gene & ASEinfo$V2==sample & !is.na(ASErawExp_pval)],1))}
else if(length(which(ASEinfo$V1==gene & ASEinfo$V2==sample & !is.na(ASErawExp_pval)))>10){
SampSigdown10 = c(SampSigdown10,median(sapply(1:10, function(q) sumlog(sample(sapply(ASErawExp_pval[ASEinfo$V1==gene & ASEinfo$V2==sample & !is.na(ASErawExp_pval)], function(x) min(x,1)), 10))[['p']])))
}
else{SampSigdown10 = c(SampSigdown10,sumlog(sapply(ASErawExp_pval[ASEinfo$V1==gene & ASEinfo$V2==sample & !is.na(ASErawExp_pval)], function(x) min(x,1)))[['p']])}
}
}
#Compute weighted ASE
ASEsW100 = c()
for(sample in samples){
testGenes = unique(ASEinfo$V1[ASEinfo$V2==sample])
ASEsW100 = c(ASEsW100, sapply(testGenes, function(gene) weighted.median(abs((Rc-ASEinfo$V4/ASEinfo$V5)[ASEinfo$V1==gene & ASEinfo$V2==sample]), sapply(ASEinfo$V7[ASEinfo$V1==gene & ASEinfo$V2==sample], function(x) min(x,100)))))
}
data.frame(sample=allSamples, gene=allTestGenes, pval=SampSigdown10, ASE=ASEsW100)
}
computeTumorSampleASE = function(ASEinfo){
#if multiple samples are being tested together
samples = unique(ASEinfo$V2)
#trim sites with no tumor reads
ASEinfo = ASEinfo[ASEinfo$V7!=0,]
#compute pvalues per heterozygous site
ASEraw_pval = c()
for(x in 1:dim(ASEinfo)[1]){
downsampledCount = min(ASEinfo$V7[x],100)
cancer_reads_raw = downsampledCount
pNormal = .5
pCancer_raw = ASEinfo$V6[x]/ASEinfo$V7[x]
ASEraw_pval = c(ASEraw_pval, 2*min(pnorm(pCancer_raw*cancer_reads_raw, pNormal*cancer_reads_raw, sqrt(cancer_reads_raw*pNormal*(1-pNormal))),pnorm(pCancer_raw*cancer_reads_raw, pNormal*cancer_reads_raw, sqrt(cancer_reads_raw*pNormal*(1-pNormal)), lower.tail = FALSE)))
}
#Combine p-values, downsampling as needed
SampSigdown10 = c()
allTestGenes = c()
allSamples = c()
for(sample in samples){
testGenes = unique(ASEinfo$V1[ASEinfo$V2==sample])
allTestGenes = c(allTestGenes, testGenes)
allSamples = c(allSamples, rep(sample, length(testGenes)))
for(gene in testGenes){
if(length(which(ASEinfo$V1==gene & ASEinfo$V2==sample))==0){SampSigdown10=c(SampSigdown10,NA)}
else if(length(which(ASEinfo$V1==gene & ASEinfo$V2==sample))==1){SampSigdown10=c(SampSigdown10,min(ASEraw_pval[ASEinfo$V1==gene & ASEinfo$V2==sample],1))}
else if(length(which(ASEinfo$V1==gene & ASEinfo$V2==sample))>10){
SampSigdown10 = c(SampSigdown10,median(sapply(1:10, function(q) sumlog(sample(sapply(ASEraw_pval[ASEinfo$V1==gene & ASEinfo$V2==sample], function(x) min(x,1)), 10))[['p']])))
}
else{SampSigdown10 = c(SampSigdown10,sumlog(sapply(ASEraw_pval[ASEinfo$V1==gene & ASEinfo$V2==sample], function(x) min(x,1)))[['p']])}
}
}
#Compute weighted ASE
ASEsW100 = c()
for(sample in samples){
testGenes = unique(ASEinfo$V1[ASEinfo$V2==sample])
ASEsW100 = c(ASEsW100, sapply(testGenes, function(gene) weighted.median(abs((ASEinfo$V6/ASEinfo$V7-.5)[ASEinfo$V1==gene & ASEinfo$V2==sample]), sapply(ASEinfo$V7[ASEinfo$V1==gene & ASEinfo$V2==sample], function(x) min(x,100)))))
}
data.frame(sample=allSamples, gene=allTestGenes, pval=SampSigdown10, ASE=ASEsW100)
}
#Simulate ASE reads at a heterozyogus site
generateReads = function(Rc, Rn=.5, p=.8, en=1, ec=1, AF=1, numReads=20, numCells=10000, errorRate=1/10000){
trueASE = abs(Rc-Rn)
cancerCells = rbinom(1, numCells, p)
mutatedCells = rbinom(1, cancerCells, min(AF/p,1))
unmutatedCells = cancerCells-mutatedCells
normalCells = numCells-cancerCells
mutatedRna = mutatedCells*ec
unmutatedRna = unmutatedCells*en
normalRna = normalCells*en
mutatedRnaMajor = rbinom(1,round(mutatedRna),Rc)
unmutatedRnaMajor = rbinom(1,round(unmutatedRna),Rn)
normalRnaMajor = rbinom(1,round(normalRna),Rn)
mutatedRnaMajor = rbinom(1,mutatedRnaMajor,1-errorRate)+rbinom(1,mutatedRna-mutatedRnaMajor,errorRate)
unmutatedRnaMajor = rbinom(1,unmutatedRnaMajor,1-errorRate)+rbinom(1,unmutatedRna-unmutatedRnaMajor,errorRate)
normalRnaMajor = rbinom(1,normalRnaMajor,1-errorRate)+rbinom(1,normalRna-normalRnaMajor,errorRate)
Rs = (mutatedRnaMajor+unmutatedRnaMajor+normalRnaMajor)/(mutatedRna+unmutatedRna+normalRna)
majorReads = rbinom(1,numReads,Rs)
normalMajorReads = rbinom(1,round(Rn*numReads),1-errorRate)+rbinom(1,numReads-round(Rn*numReads),errorRate)
c(normalMajorReads, numReads, majorReads, numReads, trueASE)
}
#simple test
testASE = data.frame(1,"sample1", "mutLoc", t(generateReads(.7, Rn=.6, p=.8, en=2)[1:4]))
colnames(testASE) = paste0("V",1:7)
aseBaseline = computeASEbaseline(testASE)
allPurity = data.frame(V1=c("sample1"), V2=c(.8))
asePurity = computeASEpurity(testASE)
cpmTumor = data.frame("sample1"=1)
rownames(cpmTumor) = "ENSG00000121410"
cpmNormal = data.frame("sample1"=2)
rownames(cpmNormal) = "ENSG00000121410"
aseExp = computeASEexp(testASE)
aseTumorSample = computeTumorSampleASE(testASE)