adjustBetas/example_tcgaBrcaRun.r at master · StaafLab/adjustBetas · 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
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
#####======================================================================#####
### Correct TCGA BRCA beta values for infiltration
#####======================================================================#####

##Author: Mattias Aine  (mattias.aine@med.lu.se)
##Affiliation: Lund University / Oncoloy and Pathology

##Git for data download and preprocessing of TCGA BRCA data (Staaf-lab pipeline)
#https://github.com/StaafLab/tcgaBrca
#resluting data object can be obtained on demand from MA/Staaf-lab to avoid extensive download/run times

##Git for beta adjustment
#https://github.com/StaafLab/adjustBetas

################################################################################
##Set directory paths

##set/create own home directory below:
HOME<-"~/hdd1/tcgaBrca"
MANIFEST<-"~/Documents/tcgaBrca/manifest"

GIT<-"~/Documents/adjustBetas"

##tumor type
TUMOR_TYPE<-"brca"

################################################################################
##load required packages

if(!requireNamespace("doParallel", quietly = TRUE)) {
  install.packages("doParallel") }

library(doParallel)

if(!requireNamespace("parallel", quietly = TRUE)) {
  install.packages("parallel") }

library(parallel)

##source - flexmix loaded on source
source(paste0(GIT,"/function_correctBetas.r"))

################################################################################
##Get data object and purity vector

load(file=paste0(HOME,"/","finalWorkspace_atacCnGexMeWes_withAnnotations.RData"))

ls()
#  [1] "adjustBeta" "annoObj"    "betaOrig"   "DATA"       "dataAtac"   "dataCn"     "dataMut"    "dataSeg"
#  [9] "gexCounts"  "gexFpkm"    "gexUq"      "GIT"        "HOME"       "MANIFEST"   "sampleAnno" "TUMOR_TYPE"

fracTum<-read.table(file=paste0(DATA,"/me/sampleTumorDnaFractionVector.txt"),header=TRUE,as.is=T,sep="\t")

str(fracTum)
# 'data.frame':   630 obs. of  2 variables:
#  $ sample            : chr  "TCGA-3C-AAAU-01" "TCGA-3C-AALI-01" "TCGA-3C-AALJ-01" "TCGA-3C-AALK-01" ...
#  $ tumor_dna_fraction: num  0.85 0.73 0.69 0.62 0.64 0.53 0.89 0.47 0.56 0.65 ...

rownames(fracTum)<-fracTum$sample

identical(rownames(fracTum),sub(".$","",colnames(betaOrig)))
#[1] TRUE

rownames(fracTum)<-colnames(betaOrig)

################################################################################
###Beta correction on top 5000 most varying probes

##filter X/Y and get top5k
betaData<-betaOrig

tmp<-fracTum$tumor_dna_fraction
names(tmp)<-rownames(fracTum)

all.equal(names(tmp),colnames(betaData))
#[1] TRUE

fracTum<-tmp
rm(tmp)

all.equal(rownames(annoObj),rownames(betaData))
#[1] TRUE

table(annoObj$chr)
#  chr1 chr10 chr11 chr12 chr13 chr14 chr15 chr16 chr17 chr18 chr19  chr2 chr20
# 41098 21206 25633 21624 10540 13204 13119 18627 24356  5280 21984 30550  9354
# chr21 chr22  chr3  chr4  chr5  chr6  chr7  chr8  chr9  chrX  chrY
#  3378  7222 22315 17646 21279 31170 25024 18194  8605  9718   242

table(annoObj$illuminaCpG_CpH_Probe)
#     cg     ch
# 418858   2510

betaData<-betaData[ !annoObj$chr %in% c("chrX","chrY") &
   annoObj$illuminaCpG_CpH_Probe == "cg"
,]

varF<-apply(betaData,1,sd)

varF<-varF >= quantile(varF,1-(5000/nrow(betaData)))

table(varF)
# varF
#  FALSE   TRUE
# 403966   5000

varF<-rownames(betaData)[varF]

##create data set and run function
testDat2<-betaData[varF,]

str(testDat2)
 # num [1:5000, 1:630] 0.341 0.057 0.012 0.022 0 0.004 0.021 0.892 0.022 0.163 ...
 # - attr(*, "dimnames")=List of 2
 #  ..$ : chr [1:5000] "cg09248054" "cg25340711" "cg06443533" "cg16601494" ...
 #  ..$ : chr [1:630] "TCGA-3C-AAAU-01A" "TCGA-3C-AALI-01A" "TCGA-3C-AALJ-01A" "TCGA-3C-AALK-01A" ...

################################################################################
###correct betas using multicore

#need to feed seed to function so fully reproducible

no_cores <- detectCores(logical = TRUE)

cat("using", no_cores-1,"cores","\n")

cl <- makeCluster(no_cores-1)
registerDoParallel(cl)

##estimated runtime is ~500 sec on 7 cores

clusterEvalQ(cl, {
  library("flexmix")
})

##add rng seed to each row and pass to each paralell instance
betaRun<-cbind(seed=1:nrow(testDat2),testDat2)
betaNames<-colnames(testDat2)

#clusterSetRNGStream(cl, 20200918) ##will not make exactly replicable..
res<-parRapply(cl = cl, betaRun, adjustBeta,purity=fracTum,snames=betaNames,seed=TRUE)

res2<-parRapply(cl = cl, betaRun, adjustBeta,purity=fracTum,snames=betaNames,seed=TRUE)

table(unlist(lapply(res,function(x) x$n.groups)))
  #  2    3
  # 20 4980

table(unlist(lapply(res,function(x) x$n.groups)),unlist(lapply(res2,function(x) x$n.groups)))
  #      2    3
  # 2   20    0
  # 3    0 4980

table( unlist(lapply(1:length(res),function(x) { all( unlist(res[[x]],use.names=FALSE) == unlist(res[[x]],use.names=FALSE) ) }) ) )
# TRUE
# 5000
table( unlist(lapply(1:length(res),function(x) { all( unlist(res[[x]],use.names=FALSE) == unlist(res2[[x]],use.names=FALSE) ) }) ) )
# TRUE
# 5000

##same results across runs.
rm(betaRun,betaNames,testDat,testDat2,res2,cl,no_cores,varF)

################################################################################
## object "res" is a list containing collected parameters of fitted lines as well as estimates of TME and pure tumor methylation

##One CpG output list entry contains following:
str(res[1])
# List of 1
#  $ cg09248054:List of 10
#   ..$ y.norm          : Named num [1:630] 0.251 0.104 0.077 0.512 0.236 0.433 0.358 0.774 0.326 0.293 ...
#   .. ..- attr(*, "names")= chr [1:630] "TCGA-3C-AAAU-01A" "TCGA-3C-AALI-01A" "TCGA-3C-AALJ-01A" "TCGA-3C-AALK-01A" ...
#   ..$ y.tum           : Named num [1:630] 0.357 0.045 0.019 0.618 0.919 0.539 1 0.88 1 0.976 ...
#   .. ..- attr(*, "names")= chr [1:630] "TCGA-3C-AAAU-01A" "TCGA-3C-AALI-01A" "TCGA-3C-AALJ-01A" "TCGA-3C-AALK-01A" ...
#   ..$ y.orig          : Named num [1:630] 0.341 0.061 0.037 0.578 0.673 0.489 0.966 0.824 0.709 0.737 ...
#   .. ..- attr(*, "names")= chr [1:630] "TCGA-3C-AAAU-01A" "TCGA-3C-AALI-01A" "TCGA-3C-AALJ-01A" "TCGA-3C-AALK-01A" ...
#   ..$ groups          : int [1:630] 3 1 1 3 2 3 2 3 2 2 ...
#   ..$ n.groups        : int 3
#   ..$ med.norm        : num 0.318
#   ..$ glob.cor        : num 0.246
#   ..$ avg.betaDiff    : num -0.14
#   ..$ model.intercepts: num [1:3] 0.065 1.011 0.542
#   ..$ model.slopes    : num [1:3] 0.058 -0.683 -0.106

##Simple way to extract matrices from object:

#adjusted tumor betas:
betaAdj<-do.call("rbind",lapply(res,function(x) x$y.tum))

str(betaAdj)
 # num [1:5000, 1:630] 0.357 0 0.009 0.018 0 0 0.018 1 0.022 0.036 ...
 # - attr(*, "dimnames")=List of 2
 #  ..$ : chr [1:5000] "cg09248054" "cg25340711" "cg06443533" "cg16601494" ...
 #  ..$ : chr [1:630] "TCGA-3C-AAAU-01A" "TCGA-3C-AALI-01A" "TCGA-3C-AALJ-01A" "TCGA-3C-AALK-01A" ...

#inferred normal
betaNormal<-do.call("rbind",lapply(res,function(x) x$y.norm))

#original beta values
betaOld<-do.call("rbind",lapply(res,function(x) x$y.orig))

q("no")
###END