-
Notifications
You must be signed in to change notification settings - Fork 5
/
prep_merge.R
358 lines (298 loc) · 12.8 KB
/
prep_merge.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
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
## Merge the RSE exon and gene objects and create the jx RSE object
## For specifying parameters
library('getopt')
## Specify parameters
spec <- matrix(c(
'bigwig_path', 'b', 1, 'character',
'Path to the directory with the bigwig files',
'jx_file', 'j', 1, 'character',
'Path to the first_pass_junctions.tsv.gz file from the cross-sample results',
'manifest_file', 'm', 1, 'character',
'Path to the manifest file used',
'wiggletools', 'w', 2, 'character',
"Path to wiggletools. If not provided, it's assumed that it is on the $PATH",
'wigToBigWig', 't', 2, 'character',
"Path to wigToBigWig. If not provided, it's assumed that it is on the $PATH",
'calculate_mean', 'c', 2, 'logical', 'Whether to calculate the AUC',
'tempdir', 'd', 2, 'character' , 'Path to a temporary directory to use. If left unspecified, it will use tempdir()',
'help' , 'h', 0, 'logical', 'Display help'
), byrow=TRUE, ncol=5)
opt <- getopt(spec)
## if help was asked for print a friendly message
## and exit with a non-zero error code
if (!is.null(opt$help)) {
cat(getopt(spec, usage=TRUE))
q(status=1)
}
## Load libraries
suppressMessages(library('SummarizedExperiment'))
suppressMessages(library('Hmisc'))
suppressMessages(library('devtools'))
## For testing
if(FALSE) {
opt <- list(
'bigwig_path' = '/dcl01/leek/data/sunghee_analysis/processed/coverage_bigwigs',
'jx_file' = '/dcl01/leek/data/sunghee_analysis/processed/cross_sample_results/first_pass_junctions.tsv.gz',
'manifest_file' = '/dcl01/leek/data/sunghee/all_s3.manifest',
'wiggletools' = 'wiggletools',
'wigToBigWig' = 'wigToBigWig',
'calculate_mean' = TRUE
)
}
## Check inputs
stopifnot(dir.exists(opt$bigwig_path))
stopifnot(file.exists(opt$jx_file))
stopifnot(file.exists(opt$manifest_file))
stopifnot(file.exists('introns_unique.Rdata'))
stopifnot(file.exists('hg38.sizes'))
## Check that outputs don't exist, to avoid overwriting
stopifnot(!file.exists('rse_gene.Rdata'))
stopifnot(!file.exists('rse_exon.Rdata'))
stopifnot(!file.exists('rse_jx.Rdata'))
## Are we on JHPCE? Print some helpful info
jhpce <- grepl('compute-', Sys.info()['nodename'])
if(jhpce & is.null(opt$wiggletools)) {
message(paste(Sys.time(), 'Note that you can use wiggletools with:
module load wiggletools/default
'))
}
if(jhpce & is.null(opt$wigToBigWig)) {
message(paste(Sys.time(), 'Note that you can use wigToBigWig with:
module load ucsctools
'))
}
## Set some defaults
if(is.null(opt$calculate_mean)) opt$calculate_mean <- TRUE
if(is.null(opt$wiggletools)) opt$wiggletools <- 'wiggletools'
if(is.null(opt$wigToBigWig)) opt$wigToBigWig <- 'wigToBigWig'
if(is.null(opt$tempdir)) opt$tempdir <- tempdir()
## Print options used
message(paste(Sys.time(), 'options used:'))
print(opt)
## Helper function for loading rse file
load_rse <- function(rse_file, type = 'exon') {
message(paste(Sys.time(), 'loading file', rse_file))
load(rse_file)
if(type == 'exon') {
return(rse_exon)
} else if (type == 'gene') {
return(rse_gene)
}
}
## Read manifest info
message(paste(Sys.time(), 'reading', opt$manifest_file))
manifest <- read.table(opt$manifest_file, sep = '\t', header = FALSE,
stringsAsFactors = FALSE, fill = TRUE)
## Get sample names from the manifest file. Note that a manifest file can
## have both paired-end and single-end data
manifest_cols <- apply(manifest, 1, function(x) { sum(x != '') })
manifest_samples <- sapply(seq_len(nrow(manifest)), function(i) {
manifest[i, manifest_cols[i]]
})
manifest$paired <- ifelse(manifest_cols == 5, TRUE, FALSE)
## Locate exon rse objects, load them, merge them and save results
exon_files <- dir('rse_temp', 'rse_exon_', full.names = TRUE)
rse_exon <- do.call(cbind, lapply(exon_files, load_rse))
## Assign paired-end info
colData(rse_exon)$paired <- manifest$paired[match(rownames(colData(rse_exon)),
manifest_samples)]
message(paste(Sys.time(), 'saving rse_exon.Rdata'))
save(rse_exon, file = 'rse_exon.Rdata')
rm(rse_exon, exon_files)
## Same for gene rse objects
gene_files <- dir('rse_temp', 'rse_gene_', full.names = TRUE)
rse_gene <- do.call(cbind, lapply(gene_files, load_rse, type = 'gene'))
colData(rse_gene)$paired <- manifest$paired[match(rownames(colData(rse_gene)),
manifest_samples)]
message(paste(Sys.time(), 'saving rse_gene.Rdata'))
save(rse_gene, file = 'rse_gene.Rdata')
## Get metadata information
metadata <- colData(rse_gene)
rm(rse_gene, gene_files)
## Calculate the mean bigwig if necessary
if(opt$calculate_mean) {
## Check that outputs don't exist
stopifnot(!file.exists('bw/mean.bw'))
stopifnot(!file.exists('bw/mean.wig'))
dir.create('bw', showWarnings = FALSE)
## Name resulting mean.bw file
outbw <- 'bw/mean.bw'
outwig <- 'bw/mean.wig'
scaleWig <- function(m) {
print(m)
paste(paste('scale', round(1e6*100*40 / m$auc, digits = 17),
file.path(opt$bigwig_path, m$bigwig_file)), collapse = ' ')
}
runCmd <- function(cmd, i = NULL) {
if(is.null(i)) {
shell_name <- '.createWig.sh'
} else {
shell_name <- paste0('.createWig_part', i, '.sh')
}
message(paste(Sys.time(), 'command used:', cmd))
cat(cmd, file = shell_name)
system(paste('sh', shell_name))
}
## Calculate mean bigwig
if(nrow(metadata) < 100) {
## Scale commands
cmd <- scaleWig(metadata)
## Calculate mean wig file
message(paste(Sys.time(), 'creating file', outwig))
cmd <- paste(opt$wiggletools, 'write', outwig, 'mean', cmd)
system.time( runCmd(cmd) )
} else {
## Define subsets to work on
sets <- cut2(seq_len(nrow(metadata)), m = 50)
meta <- split(metadata, sets)
names(meta) <- seq_len(length(meta))
## Calculate sums per subsets
system.time( tmpfiles <- mapply(function(m, i) {
cmd <- scaleWig(m)
tmpwig <- file.path(opt$tempdir, paste0('sum_part', i, '.wig'))
message(paste(Sys.time(), 'creating file', tmpwig))
cmd <- paste(opt$wiggletools, 'write', tmpwig, 'sum', cmd)
runCmd(cmd, i)
return(tmpwig)
}, meta, names(meta)) )
## Calculate final mean
cmd <- paste(opt$wiggletools, 'write', outwig, 'scale',
1/nrow(metadata), 'sum', paste(tmpfiles, collapse = ' '))
system.time( runCmd(cmd) )
## Clean up
sapply(tmpfiles, unlink)
}
## Transform to bigwig file
message(paste(Sys.time(), 'creating file', outbw))
cmd2 <- paste(opt$wigToBigWig, outwig, 'hg38.sizes', outbw)
system.time( system(cmd2) )
}
## Code for creating rse_jx
message(paste(Sys.time(), 'reading', opt$jx_file))
jx_info <- read.table(opt$jx_file, sep = '\t', header = FALSE,
stringsAsFactors = FALSE, check.names = FALSE)
colnames(jx_info) <- c('chr', 'start', 'end', 'sample_ids', 'reads')
## Create the counts matrix
message(paste(Sys.time(), 'processing count information'))
jx_info_samples <- strsplit(as.character(jx_info$sample_ids), ',')
jx_info_reads <- strsplit(as.character(jx_info$reads), ',')
stopifnot(identical(elementNROWS(jx_info_samples), elementNROWS(jx_info_reads)))
jx_info_tab <- data.frame(
jx_id = rep(seq_len(nrow(jx_info)), elementNROWS(jx_info_samples)),
sample_id = unlist(jx_info_samples),
reads = as.numeric(unlist(jx_info_reads)), stringsAsFactors = FALSE
)
rm(jx_info_samples, jx_info_reads)
message(paste(Sys.time(), 'creating junction counts table'))
## Create junction counts table
jx_counts <- matrix(0, ncol = nrow(metadata), nrow = nrow(jx_info))
colnames(jx_counts) <- rownames(metadata)
## Fill in table
for(sample in rownames(metadata)) {
## Note that the sample ids in opt$jx_file are 0-based
sampleId <- as.character(which(manifest_samples == sample) - 1)
sample_reads <- subset(jx_info_tab, sample_id == sampleId)
if(nrow(sample_reads) == 0) {
message(paste(Sys.time(), 'found no junction counts for sample',
sample))
next
}
jx_map <- match(seq_len(nrow(jx_info)), sample_reads$jx_id)
jx_counts[!is.na(jx_map), sample] <- sample_reads$reads[jx_map[!is.na(jx_map)]]
}
rm(sample, sampleId, sample_reads, jx_map, jx_info_tab)
## Create a GRanges object for the exon-exon junctions
message(paste(Sys.time(),
'forming GRanges object with exon-exon junctions information'))
jx_gr <- GRanges(seqnames = gsub('\\+|-', '', jx_info$chr), IRanges(
start = jx_info$start, end = jx_info$end), strand = ifelse(grepl('\\+',
jx_info$chr), '+', ifelse(grepl('-', jx_info$chr), '-', '*')))
chr_info <- read.table('hg38.sizes', sep = '\t',
col.names = c('chr', 'length'), stringsAsFactors = FALSE)
chrs <- chr_info$length
names(chrs) <- chr_info$chr
seqlengths(jx_gr) <- chrs[names(seqlengths(jx_gr))]
## Add recount columns that are NA here
jx_gr$junction_id <- as.character(NA)
jx_gr$found_junction_gencode_v24 <- rep(CharacterList(NA), length(jx_gr))
message(paste(Sys.time(),
'finding tx_name and gene_id based on the intron reference set'))
## Now to actual data, add the transcript names and gene ids
load('introns_unique.Rdata')
oo <- findOverlaps(jx_gr, introns_unique, type = 'equal')
stopifnot(length(unique(queryHits(oo))) == length(oo))
jx_gr$symbol <- jx_gr$gene_id <- jx_gr$tx_name <- jx_gr$gene_id_proposed <- jx_gr$symbol_proposed <- CharacterList(NA)
left_gene <- left_symbol <- right_gene <- right_symbol <- jx_gr$gene_id_proposed
## Partial overlap
message(paste(Sys.time(),
'finding gene ids and symbols based on partial matching'))
both <- countOverlaps(jx_gr, introns_unique, type = 'equal') > 0
not_both <- which(!both)
## Left
oo_left <- findOverlaps(jx_gr[not_both], introns_unique, type = 'start')
left_gene[not_both[queryHits(oo_left)]] <- introns_unique$gene_id[subjectHits(oo_left)]
left_symbol[not_both[queryHits(oo_left)]] <- introns_unique$symbol[subjectHits(oo_left)]
## Right
oo_right <- findOverlaps(jx_gr[not_both], introns_unique, type = 'end')
right_gene[not_both[queryHits(oo_right)]] <- introns_unique$gene_id[subjectHits(oo_right)]
right_symbol[not_both[queryHits(oo_right)]] <- introns_unique$symbol[subjectHits(oo_right)]
## Manually combine, since using paste(x, y, sep = '-') won't work
## for cases where x and y have different lengths
manual_c <- function(l, r) {
tmp <- merge(l, r, all = TRUE)
tmp2 <- tmp[!is.na(tmp$value), ]
## Add back if only it's NAs
missed <- unique(tmp$group)[!unique(tmp$group) %in% tmp2$group]
if(length(missed) > 0) {
tmp2 <- rbind(tmp2, subset(tmp, group %in% missed))
}
res <- CharacterList(split(as.character(tmp2$value), tmp2$group))
names(res) <- NULL
return(res)
}
message(paste(Sys.time(), 'combining left and right results'))
has_hit <- not_both[unique(c(queryHits(oo_left), queryHits(oo_right)))]
ends_hit <- not_both[intersect(queryHits(oo_left), queryHits(oo_right))]
jx_gr$gene_id_proposed[has_hit] <- manual_c(left_gene[has_hit],
right_gene[has_hit])
jx_gr$symbol_proposed[has_hit] <- manual_c(left_symbol[has_hit],
right_symbol[has_hit])
## Full match
jx_gr$gene_id_proposed[queryHits(oo)] <- jx_gr$gene_id[queryHits(oo)] <- introns_unique$gene_id[subjectHits(oo)]
jx_gr$symbol_proposed[queryHits(oo)] <- jx_gr$symbol[queryHits(oo)] <- introns_unique$symbol[subjectHits(oo)]
jx_gr$tx_name[queryHits(oo)] <- introns_unique$tx_name[subjectHits(oo)]
## See how junctions matched
print('Exon-exon junctions matching or partial matching')
print(addmargins(table('has proposed gene_id' = !any(is.na(jx_gr$gene_id_proposed)), 'exact match, gene_id' = !any(is.na(jx_gr$gene_id)))))
## Assign class
message(paste(Sys.time(), 'assigning class'))
left <- countOverlaps(jx_gr, introns_unique, type = 'start') > 0
right <- countOverlaps(jx_gr, introns_unique, type = 'end') > 0
jx_gr$class <- ifelse(both, 'annotated',
ifelse(left & right, 'exon_skip',
ifelse(left | right, 'alternative_end', 'novel')))
## Detect fusion
message(paste(Sys.time(), 'detecting gene fusions'))
find_fusions <- function(fu) {
tmp <- merge(left_gene[fu], right_gene[fu])
seq_len(length(fu))[!seq_len(length(fu)) %in% tmp$group]
}
jx_gr$class[ends_hit[find_fusions(ends_hit)]] <- 'fusion'
print('Exon-exon junctions by class')
print(table(jx_gr$class))
## Create the junctions level rse
message(paste(Sys.time(), 'creating rse_jx object'))
rse_jx <- SummarizedExperiment(assays = list('counts' = jx_counts),
colData = metadata, rowRanges = jx_gr)
message(paste(Sys.time(), 'saving rse_jx.Rdata'))
save(rse_jx, file = 'rse_jx.Rdata')
## Clean up
message(paste(Sys.time(), 'cleaning up temporary files'))
to_clean <- c(outwig, 'Gencode-v25.bed', 'count_groups.Rdata', 'hg38.sizes',
'introns_unique.Rdata')
sapply(to_clean, unlink)
unlink('rse_temp', recursive = TRUE)
## Reproducibility info
proc.time()
options(width = 120)
session_info()