-
Notifications
You must be signed in to change notification settings - Fork 4
/
qtl.metadata.R
184 lines (161 loc) · 8.25 KB
/
qtl.metadata.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
#' This is an example script which loads either eQTL or pQTL scores depending
#' on the value of `analysis` variable. In addition, this script requires that
#' the path to the scores directory is specified in the global environment.
## =============================================================================
if (analysis == "eQTL") {
## load processed scores and metadata
eqtl.datadir <- score.dir
load(file.path(eqtl.datadir, "trans.scoresinfo.1e-6.Rdata.gz"))
load(file.path(eqtl.datadir, "trans.genotypicscore.1e-6.Rdata.gz"))
## clean up metadata
trans.genome.wide.scoresinfo[, chrom_min :=
factor(chrom_min, levels=c(1:22, "X", "Y"))]
trans.genome.wide.scoresinfo[, x.startpos :=
position.absolute(chrom_min, startpos) - DMhits.flanksize.kb * 1000]
trans.genome.wide.scoresinfo[, x.endpos :=
position.absolute(chrom_min, endpos) + DMhits.flanksize.kb * 1000]
setkey(trans.genome.wide.scoresinfo, x.startpos, x.endpos)
setkey(t1d.hits, x.clumpStart, x.clumpEnd)
eqtl.overlaps <- foverlaps(trans.genome.wide.scoresinfo[, .(scoreid, x.startpos, x.endpos)],
t1d.hits[, .(nearest, x.clumpStart, x.clumpEnd)], nomatch=NULL)
eqtl.overlaps <- eqtl.overlaps[, .(near.t1dgenes=paste(nearest, collapse=",")), by=scoreid]
setkey(eqtl.overlaps, scoreid)
setkey(trans.genome.wide.scoresinfo, scoreid)
trans.genome.wide.scoresinfo <- eqtl.overlaps[trans.genome.wide.scoresinfo]
setorder(trans.genome.wide.scoresinfo, chrom_min, startpos, endpos)
transqtl.loci <- trans.genome.wide.scoresinfo[qtl_type=="trans",
.(scoreid,
gene_symbol,
chrom_min,
startpos,
endpos,
nearby_genes,
near.t1dgenes)]
transqtl.loci[, x.startpos := position.absolute(chrom_min, startpos)]
transqtl.loci[, x.endpos := position.absolute(chrom_min, endpos)]
setorder(transqtl.loci, chrom_min, x.startpos)
transqtl.loci[grep("^STH|NSF$", nearby_genes)]
## one score clump has STH included as nearby.t1dgene though it is >600 kb downstream of transcription site
## this is because SNPs in the T1D GWAS hit extend from 45.32 to 46.50 Mb
## define region by minimum gap of between last endpos and startpos
transqtl.loci[, newregion :=
c(1, as.integer((x.startpos[-1] - x.endpos[-.N] > 1E5)))]
transqtl.loci[, region := cumsum(newregion)]
setkey(transqtl.loci, scoreid)
setkey(trans.genome.wide.scoresinfo, scoreid)
trans.genome.wide.scoresinfo <-
transqtl.loci[, .(scoreid, region)][trans.genome.wide.scoresinfo]
setorder(transqtl.loci, chrom_min, startpos)
## collapse to one row per region as transqtl.regions
transqtl.regions <- transqtl.loci[,
.(chrom=chrom_min[1],
startpos=min(startpos),
endpos=max(endpos),
numscores=.N,
targetgenes=paste(unique(gene_symbol), collapse=","),
nearbygenes=paste(unique(unlist(strsplit(nearby_genes, split=","))), collapse=","),
near.t1dgenes = paste(unique(unlist(strsplit(near.t1dgenes, split=","))), collapse = ",")),
by=region]
setkey(transqtl.regions, region)
transqtl.regions[nchar(near.t1dgenes) > 0] # regions containing known T1D-associated SNPs
## generate metadata table with rows matching colnames of scores matrix:
## i.e. genomewide trans scores
## for genomewide trans scores the scores matrix colname is the gene symbol
## for cis scores the scores matrix colname is the locus-specific scoreid
tscores.info <- trans.genome.wide.scoresinfo[qtl_type=="trans",
.(numscores=.N,
matrix.colname=paste0("X_", gwasid, "_trans"),
qtl_type="trans",
gene_symbol=gene_symbol[1],
gene_chrom=gene_chrom[1],
gene_startpos=gene_startpos[1],
gene_endpos=gene_endpos[1],
regions=paste(sort(unique(region)),
collapse=","),
locus.diversity=diversity(variance, 1)),
by=.(traitid, gwasid)]
cscores.info <- trans.genome.wide.scoresinfo[substr(qtl_type, 1, 3)=="cis",
.(scoreid,
matrix.colname=scoreid,
qtl_type,
gene_symbol,
gene_chrom,
gene_startpos, gene_endpos)]
## combine the trans and cis scores
allscores.info <- rbind(tscores.info, cscores.info, fill=TRUE)
## column matrix.colname in allscores.info matches column names in scores table
## recode cis-x as cis but keep an indicator for cis-x
allscores.info[, cisx := qtl_type=="cis-x"]
allscores.info[qtl_type=="cis-x", qtl_type := "cis"]
table(allscores.info$matrix.colname %in% colnames(genome.wide.scores))
## remove rows not matched in scores table
allscores.info <- allscores.info[matrix.colname %in% colnames(genome.wide.scores)]
scores.sd <- apply(genome.wide.scores, 2, sd)
scores.sd <- data.table(matrix.colname=names(scores.sd), sdscore=as.numeric(scores.sd))
setkey(scores.sd, matrix.colname)
setkey(allscores.info, matrix.colname)
allscores.info <- scores.sd[allscores.info]
}
if (analysis == "pQTL") {
## load processed scores and metadata
pqtl.datadir <- score.dir
load(file.path(pqtl.datadir, "trans.scoresinfo.1e-6.Rdata.gz"))
load(file.path(pqtl.datadir, "trans.genotypicscore.1e-6.Rdata.gz"))
## rename and clean up metadata
## cis- columns have name X_nnn_n_n
## trans- columns have gene name
trans.genome.wide.scoresinfo[, gene_symbol := gsub("character\\(0\\),?", "", gene_symbol)]
trans.genome.wide.scoresinfo[, gene_symbol := gsub("c\\(", "", gene_symbol)]
trans.genome.wide.scoresinfo[, gene_symbol := gsub("\\)", "", gene_symbol)]
trans.genome.wide.scoresinfo[, gene_symbol := gsub('\\"', "", gene_symbol)]
trans.genome.wide.scoresinfo[, gene_symbol := gsub(" TBCE", "", gene_symbol)]
trans.genome.wide.scoresinfo[, gene_symbol := gsub(",$", "", gene_symbol)]
trans.genome.wide.scoresinfo[is.na(gene_chrom), gene_symbol]
trans.genome.wide.scoresinfo <- trans.genome.wide.scoresinfo[nchar(gene_symbol) > 0]
## FIXME: some trans scores do not map to a single gene
trans.genome.wide.scoresinfo[, gene_symbol := gsub(",.*", "", gene_symbol)]
pqtl.genes.missinginfo <- trans.genome.wide.scoresinfo[is.na(gene_chrom), gene_symbol]
missinginfo <- genes.info[match(pqtl.genes.missinginfo, gene_symbol),
.(chromosome, gene_startpos, gene_endpos)]
trans.genome.wide.scoresinfo[is.na(gene_chrom),
`:=`(gene_chrom=missinginfo$chromosome,
gene_startpos=missinginfo$gene_startpos,
gene_endpos=missinginfo$gene_endpos)]
options(warn=-1)
trans.genome.wide.scoresinfo[, gene_chrom :=
factor(as.character(gene_chrom), levels=c(1:22, "X", "Y"))]
trans.genome.wide.scoresinfo[, gene_startpos := as.integer(gene_startpos)]
trans.genome.wide.scoresinfo[, gene_endpos := as.integer(gene_endpos)]
trans.genome.wide.scoresinfo[, chrom_min :=
factor(chrom_min, levels=c(1:22, "X", "Y"))]
options(warn=2)
summary(trans.genome.wide.scoresinfo)
## trans scores identified in matrix only by gene_symbol, so may have been
## summed over multiple gwasids
pqtl.tscores.info <- unique(trans.genome.wide.scoresinfo[qtl_type=="trans",
.(numscores=.N,
scoreids=paste(scoreid, collapse=","),
gwasids=paste(gwasid, collapse=","),
numgwas=length(unique(gwasid)),
matrix.colname=paste0("X_", gwasid, "_trans"),
qtl_type="trans",
gene_symbol=gene_symbol[1],
gene_chrom=gene_chrom[1],
gene_startpos=gene_startpos[1],
gene_endpos=gene_endpos[1],
regions=paste(sort(unique(region)),
collapse=","),
locus.diversity=diversity(variance, 1)),
by=gene_symbol])
## cis scores are uniquely identified by scoreid
pqtl.cscores.info <- trans.genome.wide.scoresinfo[substr(qtl_type, 1, 3)=="cis",
.(matrix.colname=scoreid,
scoreids=paste(scoreid, collapse=","),
gwasids=gwasid,
qtl_type,
gene_symbol,
gene_chrom,
gene_startpos, gene_endpos),
by=.(traitid, gwasid)]
pqtl.allscores.info <- rbind(pqtl.tscores.info, pqtl.cscores.info, fill=TRUE)
}