-
Notifications
You must be signed in to change notification settings - Fork 0
/
estimate_HS_proportions.R
175 lines (132 loc) · 5.35 KB
/
estimate_HS_proportions.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
################################################################################
# Estimate the number of crossovers in HS mice to get an estimate of initial
# founder proportions.
# Daniel Gatti
# 2021-08-28
################################################################################
options(stringsAsFactors = FALSE)
### LIBRARIES ###
library(tidyverse)
library(qtl2convert)
library(qtl2)
### VARIABLES ###
base_dir = '/media/dmgatti/data1/ColoState/HS'
data_dir = file.path(base_dir, 'data')
data_file = file.path(data_dir, 'HZEproject_qtl2.Rdata')
results_dir = file.path(base_dir, 'results')
# NOTE: Megamuga genotypes.
muga_dir = '/media/dmgatti/data0/MUGA'
marker_file = file.path(muga_dir, 'mm_uwisc_v1.csv')
founder_names = c('A/J', 'AKR/J', 'BALB/cJ', 'C3H/HeJ', 'C57BL/6J', 'CBA/J', 'DBA/2J', 'LP/J')
HScolors = CCcolors
HScolors[1] = '#FFC800'
names(HScolors) = founder_names
### CODE ###
# Read in the big Rdata file.
load(data_file)
# Read in markers, convert to map, and synch with probs.
markers = read.csv(marker_file)[,1:3]
markers = subset(markers, chr %in% c(1:19, 'X'))
markers$bp_mm10 = markers$bp_mm10 * 1e-6
colnames(markers)[3] = 'pos'
map = map_df_to_list(markers, pos_column = 'pos')
for(i in seq_along(probs)) {
map[[i]] = map[[i]][dimnames(probs[[i]])[[3]]]
} # for(i)
# Remove objects that we don't need.
rm(addcovar, K)
# Count crossovers.
# We can use maxmarg because it needs the 36 state probs to do this correctly.
# We'll just use the autosomes.
gt = vector('list', length(probs) - 1)
names(gt) = names(probs)[-20]
for(chr in names(probs)) {
print(paste('CHR', chr))
gt[[chr]] = matrix('', nrow = nrow(probs[[chr]]), ncol = dim(probs[[chr]])[3],
dimnames = list(rownames(probs[[chr]]), dimnames(probs[[chr]])[[3]]))
pr = round(2 * probs[[chr]])
for(j in 1:dim(pr)[3]) {
x = apply(pr[,,j] == 1, 1, which)
het = which(sapply(x, length) == 2)
gt[[chr]][het, j] = apply(sapply(x[het], names), 2, paste0, collapse = '')
x = apply(pr[,,j] == 2, 1, which)
hom = which(sapply(x, length) == 1)
tmp = sapply(x[hom], names)
gt[[chr]][hom, j] = paste0(tmp, tmp)
other = which(gt[[chr]][,j] == '')
if(length(other) > 0) {
tmp = apply(apply(probs[[chr]][other,,j, drop = F], 1, rank,
ties.method = 'first') >= 7, 2, which)
gt[[chr]][other, j] = paste0(colnames(probs[[chr]])[tmp[1,]],
colnames(probs[[chr]])[tmp[2,]])
} # if(length(other) > 0)
stopifnot(all(gt[[chr]][,j] != ''))
} # for(j)
} # for(chr)
num_crossovers = vector('list', length(gt))
names(num_crossovers) = names(gt)
# Count the number of crossovers in each sample.
for(chr in names(gt)) {
curr_gt = data.frame(t(gt[[chr]]))
curr_gt = lapply(curr_gt, factor)
curr_gt = lapply(curr_gt, as.numeric)
cx = lapply(curr_gt, diff)
cx = lapply(cx, '!=', 0)
cx = lapply(cx, which)
num_crossovers[[chr]] = sapply(cx, length)
} # for(chr)
total_cx = matrix(unlist(num_crossovers), nrow = length(num_crossovers[[1]]),
ncol = length(num_crossovers),
dimnames = list(names(num_crossovers[[1]]), names(num_crossovers)))
saveRDS(total_cx, file = file.path(results_dir, 'num_crossovers.rds'))
total_cx = rowSums(total_cx[,-20])
hist(total_cx)
# Get the founder proportions.
fpr = apply(probs[[1]], c(2,3), mean) %>%
t() %>%
as.data.frame()
for(chr in names(probs)[-1]) {
print(paste('CHR', chr))
fpr = bind_rows(fpr, apply(probs[[chr]], c(2,3), mean) %>%
t() %>%
as.data.frame())
} # for(chr)
fpr = right_join(markers, rownames_to_column(fpr, var = 'marker')) %>%
filter(!is.na(chr))
colnames(fpr) = c('marker', 'chr', 'pos', founder_names)
saveRDS(fpr, file.path(results_dir, 'founder_prop_genome.rds'))
# Founder proportions along chromosomes.
fpr %>%
mutate(chr = factor(chr, levels = c(1:19, 'X'))) %>%
pivot_longer(cols = `A/J`:`LP/J`, names_to = 'founder', values_to = 'freq') %>%
ggplot() +
geom_hline(aes(yintercept = 0.125), linetype = 2, color = 'grey50') +
geom_line(aes(pos, freq, color = founder)) +
scale_color_manual(values = HScolors) +
facet_wrap(~chr, ncol = 2, dir = 'v') +
labs(title = "Founder Allele Frequency by Chromosome")
# Founder proportions by chromosome.
fpr %>%
mutate(chr = factor(chr, levels = c(1:19, 'X'))) %>%
pivot_longer(cols = `A/J`:`LP/J`, names_to = 'founder', values_to = 'freq') %>%
ggplot(aes(founder, freq, fill = founder)) +
geom_boxplot() +
scale_fill_manual(values = HScolors) +
facet_wrap(~chr)
# Genome founder proportions.
fpr %>%
mutate(chr = factor(chr, levels = c(1:19, 'X'))) %>%
pivot_longer(cols = `A/J`:`LP/J`, names_to = 'founder', values_to = 'freq') %>%
ggplot(aes(founder, freq, fill = founder)) +
geom_boxplot() +
geom_hline(aes(yintercept = 0.125), linetype = 'dashed') +
scale_fill_manual(values = HScolors)
# Get mean and median contributions from each founder across the genome.
fpr %>%
mutate(chr = factor(chr, levels = c(1:19, 'X'))) %>%
pivot_longer(cols = `A/J`:`LP/J`, names_to = 'founder', values_to = 'freq') %>%
group_by(founder) %>%
summarize(mean_prop = mean(freq, na.rm = TRUE),
median_prop = median(freq, na.rm = TRUE)) %>%
write_csv(file = file.path(results_dir, 'genome_founder_prop.csv'))