-
Notifications
You must be signed in to change notification settings - Fork 7
/
Copy pathdeAnalysis.Rmd
480 lines (410 loc) · 22.9 KB
/
deAnalysis.Rmd
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
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
---
title: "Usoskin DE analysis"
author: "Koen Van den Berge"
date: "`r Sys.Date()`"
output:
html_document:
fig_height: 7
fig_width: 7
toc: yes
code_folding: hide
toc_float: yes
---
This document uses Fanny's original simulation code in `islam_sims_fc2.Rmd' and has been adapted by Koen to adopt the new simulation framework.
```{r options, echo=FALSE, results="hide",message=FALSE, error=FALSE, include=FALSE, autodep=TRUE, warning=FALSE}
setwd("/Users/koenvandenberge/Dropbox/phdKoen/singleCell/zinbwaveZingerGithub/zinbwaveZinger/realdata/usoskin/")
knitr::opts_chunk$set(fig.align="center", cache=TRUE, error=FALSE, message=FALSE, warning=TRUE)
library(zinbwave)
library(BiocParallel)
library(doParallel)
library(Biobase)
library(edgeR)
library(scales)
library(DESeq2)
library(iCOBRA) # roc
library(limma)
library(genefilter) #filtered pvalues
library(MAST)
library(RColorBrewer)
library(knitr)
load("../../datasets/esetUsoskin.RData")
```
```{r cores}
NCORES <- 2
registerDoParallel(NCORES)
register(DoparParam())
```
```{r design}
cellType= droplevels(pData(eset)[,"Level 3"])
batch = pData(eset)[,"Picking sessions"]
counts = exprs(eset)
keep = rowSums(counts>0)>9
counts=counts[keep,]
```
# Methods
## RNA-seq methods
### edgeR
```{r edgeR}
d <- DGEList(counts)
d <- suppressWarnings(edgeR::calcNormFactors(d))
design <- model.matrix(~cellType+batch)
d <- estimateDisp(d, design)
fit <- glmFit(d,design)
L <- matrix(0,nrow=ncol(fit$coefficients),ncol=11)
rownames(L) <- colnames(fit$coefficients)
colnames(L) <- c("NF1","NF2","NF3","NF4","NF5","NP1","NP2","NP3","PEP1","PEP2","TH")
L[2:11,1] <- -1/10 #NF1 vs. others
L[2:11,2] <- c(1,rep(-1/10,9)) #NF2 vs. others
L[2:11,3] <- c(-1/10,1,rep(-1/10,8)) #NF3 vs. others
L[2:11,4] <- c(rep(-1/10,2),1,rep(-1/10,7)) #NF4 vs. others
L[2:11,5] <- c(rep(-1/10,3),1,rep(-1/10,6)) #NF5 vs. others
L[2:11,6] <- c(rep(-1/10,4),1,rep(-1/10,5)) #NP1 vs. others
L[2:11,7] <- c(rep(-1/10,5),1,rep(-1/10,4)) #NP2 vs. others
L[2:11,8] <- c(rep(-1/10,6),1,rep(-1/10,3)) #NP3 vs. others
L[2:11,9] <- c(rep(-1/10,7),1,rep(-1/10,2)) #PEP1 vs. others
L[2:11,10] <- c(rep(-1/10,8),1,rep(-1/10,1)) #PEP2 vs. others
L[2:11,11] <- c(rep(-1/10,9),1) #TH vs. others
lrtListEdgeR=list()
for(i in 1:ncol(L)) lrtListEdgeR[[i]] <- zinbwave::glmWeightedF(fit,contrast=L[,i], ZI=FALSE)
padjListEdgeR=lapply(lrtListEdgeR, function(x) p.adjust(x$table$PValue,"BH"))
deGenesEdgeR=unlist(lapply(padjListEdgeR,function(x) sum(x<.05)))
deGenesEdgeR
```
### DESeq2
```{r DESeq2}
colData <- data.frame(cellType = cellType, batch=batch)
dse <- DESeqDataSetFromMatrix(countData = counts, colData = colData, design = ~ cellType+batch)
dse <- estimateSizeFactors(dse, type="poscounts")
dse <- estimateDispersions(dse, minmu=1e-3)
dse <- nbinomWaldTest(dse, modelMatrixType="standard", betaPrior=TRUE, minmu=1e-3)
resultsNames(dse) #for building contrasts, see ?results
L=matrix(0,nrow=length(resultsNames(dse)),ncol=11)
rownames(L)=resultsNames(dse)
L[2:11,1] <- -1/10 #NF1 vs. others
L[2:11,2] <- c(1,rep(-1/10,9)) #NF2 vs. others
L[2:11,3] <- c(-1/10,1,rep(-1/10,8)) #NF3 vs. others
L[2:11,4] <- c(rep(-1/10,2),1,rep(-1/10,7)) #NF4 vs. others
L[2:11,5] <- c(rep(-1/10,3),1,rep(-1/10,6)) #NF5 vs. others
L[2:11,6] <- c(rep(-1/10,4),1,rep(-1/10,5)) #NP1 vs. others
L[2:11,7] <- c(rep(-1/10,5),1,rep(-1/10,4)) #NP2 vs. others
L[2:11,8] <- c(rep(-1/10,6),1,rep(-1/10,3)) #NP3 vs. others
L[2:11,9] <- c(rep(-1/10,7),1,rep(-1/10,2)) #PEP1 vs. others
L[2:11,10] <- c(rep(-1/10,8),1,rep(-1/10,1)) #PEP2 vs. others
L[2:11,11] <- c(rep(-1/10,9),1) #TH vs. others
resList=list()
for(i in 1:ncol(L)) resList[[i]] = results(dse,contrast=L[,i])
lapply(resList, function(x) sum(x$padj<=0.05, na.rm=TRUE))
```
### limma-voom
```{r limma}
batchLimma=factor(gsub(batch,pattern="-",replacement="")) #valid names
design <- model.matrix(~ cellType+batchLimma)
nf <- suppressWarnings(edgeR::calcNormFactors(counts))
y <- voom(counts, design, plot = FALSE, lib.size = colSums(counts) * nf)
fit <- lmFit(y, design)
contrast.matrix = makeContrasts(
NF1 = -0.1*cellTypeNF2 - 0.1*cellTypeNF3 - 0.1*cellTypeNF4 - 0.1*cellTypeNF5 - 0.1*cellTypeNP1 - 0.1*cellTypeNP2 - 0.1*cellTypeNP3 - 0.1*cellTypePEP1 - 0.1*cellTypePEP2 - 0.1*cellTypeTH,
NF2 = cellTypeNF2 - 0.1*cellTypeNF3 - 0.1*cellTypeNF4 - 0.1*cellTypeNF5 - 0.1*cellTypeNP1 - 0.1*cellTypeNP2 - 0.1*cellTypeNP3 - 0.1*cellTypePEP1 - 0.1*cellTypePEP2 - 0.1*cellTypeTH,
NF3 = -0.1*cellTypeNF2 + cellTypeNF3 - 0.1*cellTypeNF4 - 0.1*cellTypeNF5 - 0.1*cellTypeNP1 - 0.1*cellTypeNP2 - 0.1*cellTypeNP3 - 0.1*cellTypePEP1 - 0.1*cellTypePEP2 - 0.1*cellTypeTH,
NF4 = -0.1*cellTypeNF2 - 0.1*cellTypeNF3 + cellTypeNF4 - 0.1*cellTypeNF5 - 0.1*cellTypeNP1 - 0.1*cellTypeNP2 - 0.1*cellTypeNP3 - 0.1*cellTypePEP1 - 0.1*cellTypePEP2 - 0.1*cellTypeTH,
NF5 = -0.1*cellTypeNF2 - 0.1*cellTypeNF3 - 0.1*cellTypeNF4 + cellTypeNF5 - 0.1*cellTypeNP1 - 0.1*cellTypeNP2 - 0.1*cellTypeNP3 - 0.1*cellTypePEP1 - 0.1*cellTypePEP2 - 0.1*cellTypeTH,
NP1 = -0.1*cellTypeNF2 - 0.1*cellTypeNF3 - 0.1*cellTypeNF4 - 0.1*cellTypeNF5 + cellTypeNP1 - 0.1*cellTypeNP2 - 0.1*cellTypeNP3 - 0.1*cellTypePEP1 - 0.1*cellTypePEP2 - 0.1*cellTypeTH,
NP2 = -0.1*cellTypeNF2 - 0.1*cellTypeNF3 - 0.1*cellTypeNF4 - 0.1*cellTypeNF5 - 0.1*cellTypeNP1 + cellTypeNP2 - 0.1*cellTypeNP3 - 0.1*cellTypePEP1 - 0.1*cellTypePEP2 - 0.1*cellTypeTH,
NP3 = -0.1*cellTypeNF2 - 0.1*cellTypeNF3 - 0.1*cellTypeNF4 - 0.1*cellTypeNF5 - 0.1*cellTypeNP1 - 0.1*cellTypeNP2 + cellTypeNP3 - 0.1*cellTypePEP1 - 0.1*cellTypePEP2 - 0.1*cellTypeTH,
PEP1 = -0.1*cellTypeNF2 - 0.1*cellTypeNF3 - 0.1*cellTypeNF4 - 0.1*cellTypeNF5 - 0.1*cellTypeNP1 - 0.1*cellTypeNP2 - 0.1*cellTypeNP3 + cellTypePEP1 - 0.1*cellTypePEP2 - 0.1*cellTypeTH,
PEP2 = -0.1*cellTypeNF2 - 0.1*cellTypeNF3 - 0.1*cellTypeNF4 - 0.1*cellTypeNF5 - 0.1*cellTypeNP1 - 0.1*cellTypeNP2 - 0.1*cellTypeNP3 - 0.1*cellTypePEP1 + cellTypePEP2 - 0.1*cellTypeTH,
TH = -0.1*cellTypeNF2 - 0.1*cellTypeNF3 - 0.1*cellTypeNF4 - 0.1*cellTypeNF5 - 0.1*cellTypeNP1 - 0.1*cellTypeNP2 - 0.1*cellTypeNP3 - 0.1*cellTypePEP1 - 0.1*cellTypePEP2 + cellTypeTH,
levels=design
)
fit2 <- contrasts.fit(fit, contrast.matrix)
fit2 <- eBayes(fit2)
ttList = list()
for(i in 1:11) ttList[[i]] <- topTable(fit2, coef = i, n = nrow(counts), sort.by = "none")
deGenesLimma = lapply(ttList, function(x) sum(x$adj.P.Val<=0.05))
deGenesLimma
```
## scRNA-seq methods
### MAST
```{r MAST}
counts = exprs(eset)
tpm <- counts*1e6/colSums(counts)
tpm <- log2(tpm+1)
sca <- FromMatrix(tpm, cData=data.frame(cellType=cellType, batch=batch))
#sca <- FromMatrix(counts, cData=data.frame(group=group))
# Adaptive thresholding from MAST vignette
freq_expressed <- 0.1
thres <- thresholdSCRNACountMatrix(assay(sca), nbins = 200, min_per_bin = 50, conditions = cellType)
#par(mfrow=c(5,4))
#plot(thres)
assays(sca) <- list(thresh=thres$counts_threshold, tpm=assay(sca))
expressed_genes <- freq(sca) > freq_expressed
sca <- sca[expressed_genes,]
ngeneson <- apply(exprs(eset),2,function(x) mean(x>0))
CD <- colData(sca)
CD$ngeneson <- ngeneson
CD$cngeneson <- CD$ngeneson-mean(ngeneson)
colData(sca) <- CD
## differential expression
fit <- zlm(~ cellType + batch + cngeneson, sca = sca)
### contrasts
L <- matrix(0,nrow=ncol(fit@coefC),ncol=11)
rownames(L) <- colnames(fit@coefC)
colnames(L) <- c("NF1","NF2","NF3","NF4","NF5","NP1","NP2","NP3","PEP1","PEP2","TH")
L[2:11,1] <- -1/10 #NF1 vs. others
L[2:11,2] <- c(1,rep(-1/10,9)) #NF2 vs. others
L[2:11,3] <- c(-1/10,1,rep(-1/10,8)) #NF3 vs. others
L[2:11,4] <- c(rep(-1/10,2),1,rep(-1/10,7)) #NF4 vs. others
L[2:11,5] <- c(rep(-1/10,3),1,rep(-1/10,6)) #NF5 vs. others
L[2:11,6] <- c(rep(-1/10,4),1,rep(-1/10,5)) #NP1 vs. others
L[2:11,7] <- c(rep(-1/10,5),1,rep(-1/10,4)) #NP2 vs. others
L[2:11,8] <- c(rep(-1/10,6),1,rep(-1/10,3)) #NP3 vs. others
L[2:11,9] <- c(rep(-1/10,7),1,rep(-1/10,2)) #PEP1 vs. others
L[2:11,10] <- c(rep(-1/10,8),1,rep(-1/10,1)) #PEP2 vs. others
L[2:11,11] <- c(rep(-1/10,9),1) #TH vs. others
lrFitList = list()
for(i in 1:ncol(L)) lrFitList[[i]] <- lrTest(fit, L[,i,drop=FALSE])
pval = lapply(lrFitList, function(x) x[, 'hurdle', 'Pr(>Chisq)'])
padj = lapply(pval, function(x) p.adjust(x,"BH"))
unlist(lapply(padj, function(x) sum(x<=0.05)))
```
### zinbwave
```{r}
core <- SummarizedExperiment(counts,
colData = data.frame(cellType = cellType, batch=batch))
zinb_c <- zinbFit(core, X = '~ cellType + batch', commondispersion = TRUE, epsilon=1e12)
weights = computeObservationalWeights(zinb_c, counts)
```
### zinbwave-edgeR
```{r zinbwaveedger}
d <- DGEList(counts)
d <- suppressWarnings(edgeR::calcNormFactors(d))
design <- model.matrix(~cellType+batch)
d$weights = weights
d <- estimateDisp(d, design)
fit <- glmFit(d,design)
L <- matrix(0,nrow=ncol(fit$coefficients),ncol=11)
rownames(L) <- colnames(fit$coefficients)
colnames(L) <- c("NF1","NF2","NF3","NF4","NF5","NP1","NP2","NP3","PEP1","PEP2","TH")
L[2:11,1] <- -1/10 #NF1 vs. others
L[2:11,2] <- c(1,rep(-1/10,9)) #NF2 vs. others
L[2:11,3] <- c(-1/10,1,rep(-1/10,8)) #NF3 vs. others
L[2:11,4] <- c(rep(-1/10,2),1,rep(-1/10,7)) #NF4 vs. others
L[2:11,5] <- c(rep(-1/10,3),1,rep(-1/10,6)) #NF5 vs. others
L[2:11,6] <- c(rep(-1/10,4),1,rep(-1/10,5)) #NP1 vs. others
L[2:11,7] <- c(rep(-1/10,5),1,rep(-1/10,4)) #NP2 vs. others
L[2:11,8] <- c(rep(-1/10,6),1,rep(-1/10,3)) #NP3 vs. others
L[2:11,9] <- c(rep(-1/10,7),1,rep(-1/10,2)) #PEP1 vs. others
L[2:11,10] <- c(rep(-1/10,8),1,rep(-1/10,1)) #PEP2 vs. others
L[2:11,11] <- c(rep(-1/10,9),1) #TH vs. others
lrtListZinbwaveEdger=list()
for(i in 1:ncol(L)) lrtListZinbwaveEdger[[i]] <- zinbwave::glmWeightedF(fit,contrast=L[,i])
padjListZinbEdgeR=lapply(lrtListZinbwaveEdger, function(x) p.adjust(x$table$PValue,"BH"))
deGenesZinbEdgeR=unlist(lapply(padjListZinbEdgeR,function(x) sum(x<=.05)))
deGenesZinbEdgeR
```
### zinbwave-DESeq2
```{r zinbwavedeseq2}
colData <- data.frame(cellType = cellType, batch=batch)
dse <- DESeqDataSetFromMatrix(countData = counts, colData = colData, design = ~ cellType+batch)
assays(dse)[["weights"]]=weights
dse <- DESeq2::estimateSizeFactors(dse, type="poscounts")
dse <- estimateDispersions(dse,minmu=1e-3)
dse <- nbinomWaldTest(dse, modelMatrixType="standard", betaPrior=TRUE, useT=TRUE, df=rowSums(weights)-13, minmu=1e-3)
resultsNames(dse) #for building contrasts, see ?results
L=matrix(0,nrow=length(resultsNames(dse)),ncol=11)
rownames(L)=resultsNames(dse)
L[2:11,1] <- -1/10 #NF1 vs. others
L[2:11,2] <- c(1,rep(-1/10,9)) #NF2 vs. others
L[2:11,3] <- c(-1/10,1,rep(-1/10,8)) #NF3 vs. others
L[2:11,4] <- c(rep(-1/10,2),1,rep(-1/10,7)) #NF4 vs. others
L[2:11,5] <- c(rep(-1/10,3),1,rep(-1/10,6)) #NF5 vs. others
L[2:11,6] <- c(rep(-1/10,4),1,rep(-1/10,5)) #NP1 vs. others
L[2:11,7] <- c(rep(-1/10,5),1,rep(-1/10,4)) #NP2 vs. others
L[2:11,8] <- c(rep(-1/10,6),1,rep(-1/10,3)) #NP3 vs. others
L[2:11,9] <- c(rep(-1/10,7),1,rep(-1/10,2)) #PEP1 vs. others
L[2:11,10] <- c(rep(-1/10,8),1,rep(-1/10,1)) #PEP2 vs. others
L[2:11,11] <- c(rep(-1/10,9),1) #TH vs. others
resList=list()
for(i in 1:ncol(L)) resList[[i]] = results(dse,contrast=L[,i])
lapply(resList, function(x) sum(x$padj<=0.05, na.rm=TRUE))
```
### zingeR_edgeR
```{r}
d <- DGEList(counts)
d <- suppressWarnings(edgeR::calcNormFactors(d))
design <- model.matrix(~ cellType+batch)
weights <- zingeR::zeroWeightsLS(counts = d$counts, design = design, maxit = 500, normalization = "TMM", verbose = TRUE)
d$weights <- weights
d <- estimateDisp(d, design)
fit <- glmFit(d,design)
L <- matrix(0,nrow=ncol(fit$coefficients),ncol=11)
rownames(L) <- colnames(fit$coefficients)
colnames(L) <- c("NF1","NF2","NF3","NF4","NF5","NP1","NP2","NP3","PEP1","PEP2","TH")
L[2:11,1] <- -1/10 #NF1 vs. others
L[2:11,2] <- c(1,rep(-1/10,9)) #NF2 vs. others
L[2:11,3] <- c(-1/10,1,rep(-1/10,8)) #NF3 vs. others
L[2:11,4] <- c(rep(-1/10,2),1,rep(-1/10,7)) #NF4 vs. others
L[2:11,5] <- c(rep(-1/10,3),1,rep(-1/10,6)) #NF5 vs. others
L[2:11,6] <- c(rep(-1/10,4),1,rep(-1/10,5)) #NP1 vs. others
L[2:11,7] <- c(rep(-1/10,5),1,rep(-1/10,4)) #NP2 vs. others
L[2:11,8] <- c(rep(-1/10,6),1,rep(-1/10,3)) #NP3 vs. others
L[2:11,9] <- c(rep(-1/10,7),1,rep(-1/10,2)) #PEP1 vs. others
L[2:11,10] <- c(rep(-1/10,8),1,rep(-1/10,1)) #PEP2 vs. others
L[2:11,11] <- c(rep(-1/10,9),1) #TH vs. others
lrtListZingerEdgeR=list()
for(i in 1:ncol(L)) lrtListZingerEdgeR[[i]] <- zinbwave::glmWeightedF(fit,contrast=L[,i], ZI=TRUE, independentFiltering=TRUE)
padjListZingerEdgeR=lapply(lrtListZingerEdgeR, function(x) p.adjust(x$table$PValue,"BH"))
deGenesZingerEdgeR=unlist(lapply(padjListZingerEdgeR,function(x) sum(x<.05)))
```
### zingeR_DESeq2
```{r}
colData <- data.frame(cellType = cellType, batch=batch)
design <- model.matrix(~ cellType + batch)
dse <- DESeqDataSetFromMatrix(countData = counts, colData = colData, design = ~cellType+batch)
weights <- zingeR::zeroWeightsLS(counts = counts, design = design, maxit = 500, normalization = "DESeq2_poscounts", colData = colData, designFormula = ~cellType + batch, verbose = TRUE)
assays(dse)[["weights"]] <- weights
dse <- DESeq2::estimateSizeFactors(dse, type="poscounts")
dse <- estimateDispersions(dse)
dse <- nbinomWaldTest(dse, modelMatrixType="standard", betaPrior=TRUE, useT=TRUE, df=rowSums(weights)-13, minmu=1e-3)
resultsNames(dse) #for building contrasts, see ?results
L=matrix(0,nrow=length(resultsNames(dse)),ncol=11)
rownames(L)=resultsNames(dse)
L[2:11,1] <- -1/10 #NF1 vs. others
L[2:11,2] <- c(1,rep(-1/10,9)) #NF2 vs. others
L[2:11,3] <- c(-1/10,1,rep(-1/10,8)) #NF3 vs. others
L[2:11,4] <- c(rep(-1/10,2),1,rep(-1/10,7)) #NF4 vs. others
L[2:11,5] <- c(rep(-1/10,3),1,rep(-1/10,6)) #NF5 vs. others
L[2:11,6] <- c(rep(-1/10,4),1,rep(-1/10,5)) #NP1 vs. others
L[2:11,7] <- c(rep(-1/10,5),1,rep(-1/10,4)) #NP2 vs. others
L[2:11,8] <- c(rep(-1/10,6),1,rep(-1/10,3)) #NP3 vs. others
L[2:11,9] <- c(rep(-1/10,7),1,rep(-1/10,2)) #PEP1 vs. others
L[2:11,10] <- c(rep(-1/10,8),1,rep(-1/10,1)) #PEP2 vs. others
L[2:11,11] <- c(rep(-1/10,9),1) #TH vs. others
resList=list()
for(i in 1:ncol(L)) resList[[i]] = results(dse,contrast=L[,i])
lapply(resList, function(x) sum(x$padj<=0.05, na.rm=TRUE))
```
## plot results
```{r resultsData}
results = data.frame(
nrGenes = c(46 , 106 , 3 , 37 , 105 , 415 , 2 , 3 , 54 , 1 , 381, #scde
396 , 272 , 529 , 394 , 409 , 1355 , 590 , 241 , 424 , 370 , 1425, #MAST
6944 , 5128 , 5154 , 801 , 4021 , 2422 , 6724 , 202 , 9388 , 153 , 5735, #limma
640 , 711 , 186 , 289 , 589 , 1356 , 530 , 316 , 1691 , 75 , 1652, #DESeq2
522 , 1011 , 348 , 419 , 1216 , 1915 , 762 , 376 , 2203 , 142 , 3087, #ZINB deseq2
860 , 1375 , 401 , 654 , 1149 , 1942 , 933 , 380 , 1909 , 403 , 2458, #egdeR
495 , 772 , 241 , 508 , 959 , 1802 , 967 , 464 , 2376 , 193 , 2919), #zinb edger
contrast=rep(c("NF1","NF2","NF3","NF4","NF5","NP1","NP2","NP3","PEP1","PEP2","TH"),7),
method=rep(c("SCDE","MAST","limma-voom","DESeq2","ZINB-WaVE_DESeq2", "edgeR","ZINB-WaVE_edgeR"),each=11))
results$method=factor(results$method,levels=c("SCDE","MAST","DESeq2","ZINB-WaVE_DESeq2","edgeR","ZINB-WaVE_edgeR","limma-voom"))
# library(lattice) ; library(grid)
# barchart(log10(nrGenes)~contrast,groups=method,data=results, col=c("grey","darkturquoise","brown","darkseagreen","red","salmon","blue"), page = function(n)
# grid.text(c("SCDE","MAST", "DESeq2", "ZINB-DESeq2", "edgeR", "ZINB-edgeR", "limma-voom"),
# x = seq(0.06,0.9,length=7), y = rep(.9,2),
# default.units = "npc"
# ,just = c("left", "top")))
library(ggplot2)
# Grouped
ggplot(results, aes(fill=method, y=log(nrGenes), x=contrast)) + geom_bar(position="dodge", stat="identity") + scale_fill_manual(values = c("grey","darkturquoise","brown","darkseagreen","red","salmon","blue") )+ theme_bw()
# facets
ggplot(results, aes(fill=method, y=log10(nrGenes), x=1)) + geom_bar(position="dodge", stat="identity") + facet_wrap(~contrast) + scale_fill_manual(values = c("grey","darkturquoise","brown","darkseagreen","red","salmon","blue") ) + theme_bw() + theme(axis.title.x=element_blank(),
axis.text.x=element_blank(),
axis.ticks.x=element_blank()) + ylab("log10(number of significant genes)")
### including zingeR
resultsZingeR = data.frame(
nrGenes = c(46 , 106 , 3 , 37 , 105 , 415 , 2 , 3 , 54 , 1 , 381, #scde
396 , 272 , 529 , 394 , 409 , 1355 , 590 , 241 , 424 , 370 , 1425, #MAST
6944 , 5128 , 5154 , 801 , 4021 , 2422 , 6724 , 202 , 9388 , 153 , 5735, #limma
640 , 711 , 186 , 289 , 589 , 1356 , 530 , 316 , 1691 , 75 , 1652, #DESeq2
522 , 1011 , 348 , 419 , 1216 , 1915 , 762 , 376 , 2203 , 142 , 3087, #ZINB deseq2
823, 980, 289, 461, 907, 2044, 777, 574, 1933, 145, 2404, #zingeR DESeq2
860 , 1375 , 401 , 654 , 1149 , 1942 , 933 , 380 , 1909 , 403 , 2458, #egdeR
495 , 772 , 241 , 508 , 959 , 1802 , 967 , 464 , 2376 , 193 , 2919, #zinb edgeR
981, 1740, 531, 797, 1542, 2237, 1092, 547, 2053, 492, 2769), #zingeR edger
contrast=rep(c("NF1","NF2","NF3","NF4","NF5","NP1","NP2","NP3","PEP1","PEP2","TH"),9),
method=rep(c("SCDE","MAST","limma-voom","DESeq2","ZINB-WaVE_DESeq2","zingeR_DESeq2", "edgeR","ZINB-WaVE_edgeR","zingeR_edgeR"),each=11))
resultsZingeR$method=factor(resultsZingeR$method,levels=c("SCDE","MAST","DESeq2","ZINB-WaVE_DESeq2","zingeR_DESeq2","edgeR","ZINB-WaVE_edgeR","zingeR_edgeR","limma-voom"))
# facets
ggplot(resultsZingeR, aes(fill=method, y=log10(nrGenes), x=1)) + geom_bar(position="dodge", stat="identity") + facet_wrap(~contrast) + scale_fill_manual(values = c("grey","darkturquoise","brown","darkseagreen","darkolivegreen4","red","salmon","hotpink1","blue") ) + theme_bw() + theme(axis.title.x=element_blank(),
axis.text.x=element_blank(),
axis.ticks.x=element_blank()) + ylab("log10(number of significant genes)")
```
### compare zingeR-edgeR with edgeR DE genes
```{r}
mains = c("NF1", "NF2", "NF3", "NF4", "NF5", "NP1", "NP2", "NP3", "PEP1", "PEP2", "TH")
plotVolcanoEdgeRZingeR <- function(comparison){
plot(x=lrtListZingerEdgeR[[comparison]]$table$logFC, y=-log10(lrtListZingerEdgeR[[comparison]]$table$PValue), xlab="log2 fold change", ylab="- log10(p-value)", main=mains[comparison])
padjZingeR = p.adjust(lrtListZingerEdgeR[[comparison]]$table$PValue, "BH")
padjEdgeR = p.adjust(lrtListEdgeR[[comparison]]$table$PValue, "BH")
colsBoth = (padjZingeR<=0.05 & padjEdgeR<=0.05)
colsOnlyZinger=(padjZingeR<=0.05 & padjEdgeR>0.05)
points(x=lrtListZingerEdgeR[[comparison]]$table$logFC[colsBoth], y=-log10(lrtListZingerEdgeR[[comparison]]$table$PValue[colsBoth]), col="blue", pch=16)
points(x=lrtListZingerEdgeR[[comparison]]$table$logFC[colsOnlyZinger], y=-log10(lrtListZingerEdgeR[[comparison]]$table$PValue[colsOnlyZinger]), col="red", pch=16)
legend("topleft", c("Significant in edgeR and zingeR_edgeR", "Only significant in zingeR_edgeR", "Non-significant"), pch=c(16,16,1), col=c("blue", "red", "black"), bty="n", cex=1/2)
}
png("~/Dropbox/phdKoen/singleCell/zinbwaveZinger/plots2/volcanoUsoskinZingeR.png", width=9,height=7, units="in", res=300)
library(rafalib) ; mypar()
par(mfrow=c(4,3))
sapply(1:11, plotVolcanoEdgeRZingeR)
dev.off()
plotVolcanoEdgeRZinbwave <- function(comparison){
plot(x=lrtListZinbwaveEdger[[comparison]]$table$logFC, y=-log10(lrtListZinbwaveEdger[[comparison]]$table$PValue), xlab="log2 fold change", ylab="- log10(p-value)", main=mains[comparison])
padjZinbwave = p.adjust(lrtListZinbwaveEdger[[comparison]]$table$PValue, "BH")
padjEdgeR = p.adjust(lrtListEdgeR[[comparison]]$table$PValue, "BH")
colsBoth = (padjZinbwave<=0.05 & padjEdgeR<=0.05)
colsOnlyZinger=(padjZinbwave<=0.05 & padjEdgeR>0.05)
points(x=lrtListZinbwaveEdger[[comparison]]$table$logFC[colsBoth], y=-log10(lrtListZinbwaveEdger[[comparison]]$table$PValue[colsBoth]), col="blue", pch=16)
points(x=lrtListZinbwaveEdger[[comparison]]$table$logFC[colsOnlyZinger], y=-log10(lrtListZinbwaveEdger[[comparison]]$table$PValue[colsOnlyZinger]), col="red", pch=16)
legend("topleft", c("Significant in edgeR and ZINBWaVE_edgeR", "Only significant in ZINBWaVE_edgeR", "Non-significant"), pch=c(16,16,1), col=c("blue", "red", "black"), bty="n", cex=1/2)
}
png("~/Dropbox/phdKoen/singleCell/zinbwaveZinger/plots2/volcanoUsoskinZinbwave.png", width=9,height=7, units="in", res=300)
library(rafalib) ; mypar()
par(mfrow=c(4,3))
sapply(1:11, plotVolcanoEdgeRZinbwave)
dev.off()
plot(x=lrtListZinbwaveEdger[[comparison]]$table$logFC, y=-log10(lrtListZinbwaveEdger[[comparison]]$table$PValue), xlab="log2 fold change", ylab="- log10(p-value)", main=mains[comparison])
padjZingeR = p.adjust(lrtListZinbwaveEdger[[comparison]]$table$PValue, "BH")
padjEdgeR = p.adjust(lrtListEdgeR[[comparison]]$table$PValue, "BH")
colsBoth = (padjZingeR<=0.05 & padjEdgeR<=0.05)
colsOnlyZinger=(padjZingeR<=0.05 & padjEdgeR>0.05)
points(x=lrtListZinbwaveEdger[[comparison]]$table$logFC[colsBoth], y=-log10(lrtListZinbwaveEdger[[comparison]]$table$PValue[colsBoth]), col="blue", pch=16)
points(x=lrtListZinbwaveEdger[[comparison]]$table$logFC[colsOnlyZinger], y=-log10(lrtListZinbwaveEdger[[comparison]]$table$PValue[colsOnlyZinger]), col="red", pch=16)
legend("topleft", c("Significant in edgeR and ZINBWaVE_edgeR", "Only significant in ZINBWaVE_edgeR", "Non-significant"), pch=c(16,16,1), col=c("blue", "red", "black"), bty="n", cex=1/2)
```
### Check effect of epsilon for ZINB-WaVE edgeR results
```{r}
core <- SummarizedExperiment(counts,
colData = data.frame(cellType = cellType, batch=batch))
eps = c(10^-2, 1, 1e2, 1e4, 1e6, 1e8, 1e10, 1e12, 1e14)
epsResultsDE = list()
for(i in 1:length(eps)){
zinbEps <- zinbFit(core, X = '~ cellType + batch', commondispersion = TRUE, epsilon=eps[i])
wEps = computeObservationalWeights(zinbEps, counts)
wEps[wEps<1e-6]=1e-6
d <- DGEList(counts)
d <- suppressWarnings(edgeR::calcNormFactors(d))
design <- model.matrix(~cellType+batch)
d$weights = wEps
d <- estimateDisp(d, design)
fit <- glmFit(d,design)
L <- matrix(0,nrow=ncol(fit$coefficients),ncol=11)
rownames(L) <- colnames(fit$coefficients)
colnames(L) <- c("NF1","NF2","NF3","NF4","NF5","NP1","NP2","NP3","PEP1","PEP2","TH")
L[2:11,1] <- -1/10 #NF1 vs. others
L[2:11,2] <- c(1,rep(-1/10,9)) #NF2 vs. others
L[2:11,3] <- c(-1/10,1,rep(-1/10,8)) #NF3 vs. others
L[2:11,4] <- c(rep(-1/10,2),1,rep(-1/10,7)) #NF4 vs. others
L[2:11,5] <- c(rep(-1/10,3),1,rep(-1/10,6)) #NF5 vs. others
L[2:11,6] <- c(rep(-1/10,4),1,rep(-1/10,5)) #NP1 vs. others
L[2:11,7] <- c(rep(-1/10,5),1,rep(-1/10,4)) #NP2 vs. others
L[2:11,8] <- c(rep(-1/10,6),1,rep(-1/10,3)) #NP3 vs. others
L[2:11,9] <- c(rep(-1/10,7),1,rep(-1/10,2)) #PEP1 vs. others
L[2:11,10] <- c(rep(-1/10,8),1,rep(-1/10,1)) #PEP2 vs. others
L[2:11,11] <- c(rep(-1/10,9),1) #TH vs. others
lrtList=list()
for(i in 1:ncol(L)) lrtList[[i]] <- zinbwave::glmWeightedF(fit,contrast=L[,i])
padjListZinbEdgeR=lapply(lrtList, function(x) p.adjust(x$table$PValue,"BH"))
deGenesZinbEdgeR=unlist(lapply(padjListZinbEdgeR,function(x) sum(x<=.05)))
epsResultsDE[[i]] = deGenesZinbEdgeR
}
```