-
Notifications
You must be signed in to change notification settings - Fork 0
/
gam_RR_seq_mobility.R
373 lines (322 loc) · 24.2 KB
/
gam_RR_seq_mobility.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
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
## This script implements GAM to evaluate the percentage of the variance in the genetic data
## explained by genetic data. We also identify (and plot) outliers in the relationship between
## genetic and mobility from the GAM.
library(mgcv)
library(tidyverse)
library(vegan)
library(broom)
library(ggrepel)
library(ggpubr)
source('../utils_comp_RR.R')
## Load relative risk of observing identical sequences between counties
df_RR_counties <- read_csv('../../results/RR_county/df_RR_county_0_mut_away.csv') %>% rename(RR_seq = RR)
df_RR_uncertainty_counties <- readRDS('../../results/RR_county/df_RR_uncertainty_county_0_mut_away.rds')
df_RR_regions <- read_csv('../../results/RR_region/df_RR_region_0_mut_away.csv') %>% rename(RR_seq = RR)
df_RR_uncertainty_regions <- readRDS('../../results/RR_region/df_RR_uncertainty_region_0_mut_away.rds')
## Load relative risk of movements between counties
df_RR_mobility_commute <- readRDS('../../results/RR_mobility/RR_workflow_county_WA.rds') %>% rename(RR_workflow = RR)
df_RR_mobility_commute_region <- readRDS('../../results/RR_mobility/RR_workflow_region_WA.rds') %>% rename(RR_workflow = RR)
df_RR_mobility_mobile_phone <- readRDS('../../results/RR_mobility/RR_mobile_phone_county_WA.rds') %>% rename(RR_mobile_phone = RR)
df_RR_mobility_mobile_phone_region <- readRDS('../../results/RR_mobility/RR_mobile_phone_region_WA.rds') %>% rename(RR_mobile_phone = RR)
df_distance <- readRDS('../../data/maps/df_dist_county.rds')
df_distance_region <- readRDS('../../data/maps/df_dist_region.rds')
## Join RR of identical sequences with RR of movement at the county level
df_RR_for_comparison_counties <- df_RR_counties %>% select(group_1, group_2, RR_seq, n_pairs) %>%
left_join(df_RR_mobility_commute %>% select(county_1, county_2, RR_workflow),
by = c('group_1' = 'county_1', 'group_2' = 'county_2')) %>%
left_join(df_RR_mobility_mobile_phone %>% select(county_1, county_2, RR_mobile_phone),
by = c('group_1' = 'county_1', 'group_2' = 'county_2')) %>%
left_join(df_distance, by = c('group_1' = 'county_1', 'group_2' = 'county_2')) %>%
mutate(log_RR_seq = log(RR_seq), log_RR_workflow = log(RR_workflow), log_RR_mobile_phone = log(RR_mobile_phone)) %>%
filter(group_1 >= group_2)
## Join RR of identical sequences with RR of movement at the region level
df_RR_for_comparison_regions <- df_RR_regions %>% select(group_1, group_2, RR_seq, n_pairs) %>%
left_join(df_RR_mobility_commute_region %>% select(region_1, region_2, RR_workflow),
by = c('group_1' = 'region_1', 'group_2' = 'region_2')) %>%
left_join(df_RR_mobility_mobile_phone_region %>% select(region_1, region_2, RR_mobile_phone),
by = c('group_1' = 'region_1', 'group_2' = 'region_2')) %>%
left_join(df_distance_region, by = c('group_1' = 'region_1', 'group_2' = 'region_2')) %>%
mutate(log_RR_seq = log(RR_seq), log_RR_workflow = log(RR_workflow), log_RR_mobile_phone = log(RR_mobile_phone)) %>%
filter(group_1 >= group_2)
## Load adjacency between counties
df_adj_county <- readRDS('../../data/maps/df_adj_county.rds') %>% as_tibble() %>%
rename(group_1 = county_1, group_2 = county_2)
###########################################################################
# RELATIONSHIP BETWEEN RR OF ID SEQ AND RR OF COMMUTE AT THE COUNTY LEVEL #
###########################################################################
## Run GAM and get percentage of variance explained
gam_seq_workflow_counties <- run_gam(df_with_RR_for_gam = df_RR_for_comparison_counties %>% filter(RR_workflow > 0., RR_seq > 0.),
predictor_name = 'log_RR_workflow', response_name = 'log_RR_seq',
k_gam = 5)
summary(gam_seq_workflow_counties$mod)$r.sq
summary(gam_seq_workflow_counties$mod)$s.pv
## Display the fit
plt_fit_gam_seq_workflow_counties <- plot_fit_gam_counties(res_gam = gam_seq_workflow_counties,
transform_func_response = 'exp', transform_func_predictor = 'exp',
name_x_axis = expression(RR['mobility']),
breaks_x_axis = c(1e-3, 1e-2, 1e-1, 1., 1e1, 1e2, 1e3, 1e4),
labels_x_axis = c(expression(10^{-3}), expression(10^{-2}), expression(10^{-1}),
expression(10^{0}), expression(10^{1}), expression(10^{2}), expression(10^{3}),
expression(10^{4})),
trans_x_axis_plot = 'log',
name_y_axis = expression(RR['identical sequences']),
breaks_y_axis = c(1e-3, 1e-2, 1e-1, 1., 1e1, 1e2, 1e3, 1e4),
labels_y_axis = c(expression(10^{-3}), expression(10^{-2}), expression(10^{-1}),
expression(10^{0}), expression(10^{1}), expression(10^{2}), expression(10^{3}),
expression(10^{4})),
trans_y_axis_plot = 'log',
df_adj = df_adj_county)
plt_fit_workflow_counties_with_R2 <- plt_fit_gam_seq_workflow_counties +
annotate('text', x = 1e-1, y = Inf, hjust = 1, vjust = 1.,
label = paste0('R^{2}~`=`~',
round(summary(gam_seq_workflow_counties$mod)$r.sq, 2)),
parse = T, size = 5)
## Get outliers in the relationship between genetic and mobility data
df_outliers_workflow_mobility <- get_outliers_fit(res_gam = gam_seq_workflow_counties, min_n_pairs = 100, threshold_residuals_outliers = 3.)
df_outliers_workflow_mobility
## Display outliers
plt_outliers_workflow_mobility_counties <- plot_outliers_fit_counties(res_gam = gam_seq_workflow_counties, min_n_pairs = 100, threshold_residuals_outliers = 3.)
##############################################################################################
# RELATIONSHIP BETWEEN RR OF ID SEQ AND RR OF MOVEMENT FROM MOBILE PHONE AT THE COUNTY LEVEL #
##############################################################################################
## Run GAM and get percentage of variance explained
gam_seq_mobile_phone_counties <- run_gam(df_with_RR_for_gam = df_RR_for_comparison_counties %>% filter(RR_mobile_phone > 0., RR_seq > 0.),
predictor_name = 'log_RR_mobile_phone', response_name = 'log_RR_seq',
k_gam = 5)
summary(gam_seq_mobile_phone_counties$mod)$r.sq
summary(gam_seq_mobile_phone_counties$mod)$s.pv
## Display the fit
plt_fit_gam_seq_mobile_phone_counties <- plot_fit_gam_counties(res_gam = gam_seq_mobile_phone_counties,
transform_func_response = 'exp', transform_func_predictor = 'exp',
name_x_axis = expression(RR['mobility']),
breaks_x_axis = c(1e-3, 1e-2, 1e-1, 1., 1e1, 1e2, 1e3, 1e4),
labels_x_axis = c(expression(10^{-3}), expression(10^{-2}), expression(10^{-1}),
expression(10^{0}), expression(10^{1}), expression(10^{2}), expression(10^{3}),
expression(10^{4})),
trans_x_axis_plot = 'log',
name_y_axis = expression(RR['identical sequences']),
breaks_y_axis = c(1e-3, 1e-2, 1e-1, 1., 1e1, 1e2, 1e3, 1e4),
labels_y_axis = c(expression(10^{-3}), expression(10^{-2}), expression(10^{-1}),
expression(10^{0}), expression(10^{1}), expression(10^{2}), expression(10^{3}),
expression(10^{4})),
trans_y_axis_plot = 'log',
df_adj = df_adj_county)
plt_fit_mobile_phone_counties_with_R2 <- plt_fit_gam_seq_mobile_phone_counties +
annotate('text', x = 1., y = Inf, hjust = 1, vjust = 1.,
label = paste0('R^{2}~`=`~',
round(summary(gam_seq_mobile_phone_counties$mod)$r.sq, 2)),
parse = T, size = 5)
## Get outliers in the relationship between genetic and mobility data
df_outliers_mobile_phone_mobility <- get_outliers_fit(res_gam = gam_seq_mobile_phone_counties, min_n_pairs = 100, threshold_residuals_outliers = 3.)
df_outliers_mobile_phone_mobility
## Display outliers
plt_outliers_mobile_phone_mobility_counties <- plot_outliers_fit_counties(res_gam = gam_seq_mobile_phone_counties,
min_n_pairs = 100, threshold_residuals_outliers = 3.)
######################################################################
# RELATIONSHIP BETWEEN RR OF ID SEQ AND DISTANCE AT THE COUNTY LEVEL #
######################################################################
gam_seq_distance_counties <- run_gam(df_with_RR_for_gam = df_RR_for_comparison_counties %>% filter(RR_seq > 0.),
predictor_name = 'distance_km', response_name = 'log_RR_seq',
k_gam = 5)
summary(gam_seq_distance_counties$mod)$r.sq
summary(gam_seq_distance_counties$mod)$s.pv
## Display the fit
plt_fit_gam_seq_distance_counties <- plot_fit_gam_counties(res_gam = gam_seq_distance_counties,
transform_func_response = 'exp', transform_func_predictor = 'identity',
name_x_axis = 'Distance (in km)',
breaks_x_axis = seq(0., 600., 100.),
labels_x_axis = seq(0., 600., 100.),
trans_x_axis_plot = 'identity',
name_y_axis = expression(RR['identical sequences']),
breaks_y_axis = c(1e-3, 1e-2, 1e-1, 1., 1e1, 1e2, 1e3, 1e4),
labels_y_axis = c(expression(10^{-3}), expression(10^{-2}), expression(10^{-1}),
expression(10^{0}), expression(10^{1}), expression(10^{2}), expression(10^{3}),
expression(10^{4})),
trans_y_axis_plot = 'log',
df_adj = df_adj_county)
plt_fit_distance_counties_with_R2 <- plt_fit_gam_seq_distance_counties +
annotate('text', x = 200, y = Inf, hjust = 1, vjust = 1.,
label = paste0('R^{2}~`=`~',
round(summary(gam_seq_distance_counties$mod)$r.sq, 2)),
parse = T, size = 5)
## Get outliers in the relationship between genetic and mobility data
df_outliers_distance_mobility <- get_outliers_fit(res_gam = gam_seq_distance_counties, min_n_pairs = 100, threshold_residuals_outliers = 3.)
df_outliers_distance_mobility
## Display outliers
plt_outliers_distance_counties <- plot_outliers_fit_counties(res_gam = gam_seq_distance_counties,
min_n_pairs = 100, threshold_residuals_outliers = 3.)
###########################################################################
# RELATIONSHIP BETWEEN RR OF ID SEQ AND RR OF COMMUTE AT THE REGION LEVEL #
###########################################################################
## Run GAM and get percentage of variance explained
gam_seq_workflow_regions <- run_gam(df_with_RR_for_gam = df_RR_for_comparison_regions %>% filter(RR_workflow > 0., RR_seq > 0.),
predictor_name = 'log_RR_workflow', response_name = 'log_RR_seq',
k_gam = 5)
summary(gam_seq_workflow_regions$mod)$r.sq
summary(gam_seq_workflow_regions$mod)$s.pv
## Display the fit
plt_fit_workflow_regions <- plot_fit_gam_regions(res_gam = gam_seq_workflow_regions,
transform_func_response = 'exp', transform_func_predictor = 'exp',
name_x_axis = expression(RR['mobility']),
breaks_x_axis = c(1e-3, 1e-2, 1e-1, 1., 1e1, 1e2, 1e3, 1e4),
labels_x_axis = c(expression(10^{-3}), expression(10^{-2}), expression(10^{-1}),
expression(10^{0}), expression(10^{1}), expression(10^{2}), expression(10^{3}),
expression(10^{4})),
trans_x_axis_plot = 'log',
name_y_axis = expression(RR['identical sequences']),
breaks_y_axis = c(1, 2, 5),
labels_y_axis =c(1, 2, 5),
trans_y_axis_plot = 'log')
plt_fit_workflow_regions_with_R2 <- plt_fit_workflow_regions +
annotate('text', x = 1e-1, y = Inf, hjust = 1, vjust = 1.,
label = paste0('R^{2}~`=`~',
round(summary(gam_seq_workflow_regions$mod)$r.sq, 2)),
parse = T, size = 5)
##############################################################################################
# RELATIONSHIP BETWEEN RR OF ID SEQ AND RR OF MOVEMENT FROM MOBILE PHONE AT THE REGION LEVEL #
##############################################################################################
## Run GAM and get percentage of variance explained
gam_seq_mobile_phone_regions <- run_gam(df_with_RR_for_gam = df_RR_for_comparison_regions %>% filter(RR_mobile_phone > 0., RR_seq > 0.),
predictor_name = 'log_RR_mobile_phone', response_name = 'log_RR_seq',
k_gam = 5)
summary(gam_seq_mobile_phone_regions$mod)$r.sq
summary(gam_seq_mobile_phone_regions$mod)$s.pv
## Display the fit
plt_fit_mobile_phone_regions <- plot_fit_gam_regions(res_gam = gam_seq_mobile_phone_regions,
transform_func_response = 'exp', transform_func_predictor = 'exp',
name_x_axis = expression(RR['mobility']),
breaks_x_axis = c(1e-3, 1e-2, 1e-1, 1., 1e1, 1e2, 1e3, 1e4),
labels_x_axis = c(expression(10^{-3}), expression(10^{-2}), expression(10^{-1}),
expression(10^{0}), expression(10^{1}), expression(10^{2}), expression(10^{3}),
expression(10^{4})),
trans_x_axis_plot = 'log',
name_y_axis = expression(RR['identical sequences']),
breaks_y_axis = c(1, 2, 5),
labels_y_axis =c(1, 2, 5),
trans_y_axis_plot = 'log')
plt_fit_mobile_phone_regions_with_R2 <- plt_fit_mobile_phone_regions +
annotate('text', x = 1, y = Inf, hjust = 1, vjust = 1.,
label = paste0('R^{2}~`=`~',
round(summary(gam_seq_workflow_regions$mod)$r.sq, 2)),
parse = T, size = 5)
######################################################################
# RELATIONSHIP BETWEEN RR OF ID SEQ AND DISTANCE AT THE REGION LEVEL #
######################################################################
## Run GAM and get percentage of variance explained
gam_seq_distance_regions <- run_gam(df_with_RR_for_gam = df_RR_for_comparison_regions %>% filter(RR_seq > 0.),
predictor_name = 'distance_km', response_name = 'log_RR_seq',
k_gam = 5)
summary(gam_seq_distance_regions$mod)$r.sq
summary(gam_seq_distance_regions$mod)$s.pv
## Display the fit
plt_fit_distance_regions <- plot_fit_gam_regions(res_gam = gam_seq_distance_regions,
transform_func_response = 'exp', transform_func_predictor = 'identity',
name_x_axis = 'Distance (in km)',
breaks_x_axis = seq(0., 600., 100.),
labels_x_axis = seq(0., 600., 100.),
trans_x_axis_plot = 'identity',
name_y_axis = expression(RR['identical sequences']),
breaks_y_axis = c(1, 2, 5),
labels_y_axis =c(1, 2, 5),
trans_y_axis_plot = 'log')
plt_fit_distance_regions_with_R2 <- plt_fit_distance_regions +
annotate('text', x = 200, y = Inf, hjust = 1, vjust = 1.,
label = paste0('R^{2}~`=`~',
round(summary(gam_seq_distance_regions$mod)$r.sq, 2)),
parse = T, size = 5)
#################################################################
# SCATTERPLOT OF THE RESIDUALS OF THE GAM AT THE REGIONAL LEVEL #
#################################################################
plt_residuals_regions <- gam_seq_workflow_regions$df_pred %>% mutate(type = 'Commuting') %>%
bind_rows(gam_seq_mobile_phone_regions$df_pred %>% mutate(type = 'Mobile phone')) %>%
bind_rows(gam_seq_distance_regions$df_pred %>% mutate(type = 'Distance')) %>%
mutate(type = factor(type, levels = c('Mobile phone', 'Commuting', 'Distance'))) %>%
ggplot(aes(x = type, y = scaled_pearson_resid, colour = (group_1 == group_2))) +
geom_jitter(width = 0.2) +
scale_x_discrete(name = '') +
scale_colour_manual(name = '', breaks = c(T, F), values = c('firebrick2', 'black'),
labels = c('Within region', 'Between regions')) +
scale_y_continuous(name = 'Scaled Pearson residuals', breaks = seq(-4, 4, 1),
limits = c(-4, 4)) +
theme_classic() +
theme(axis.text = element_text(size = 12),
axis.title = element_text(size = 12),
legend.text = element_text(size = 12),
legend.background = element_blank(),
legend.position = c(0.3, 0.95))
##################
## SAVE FIGURES ##
##################
## GAM between RR of id seq and RR of movement from mobile phone data (Figure 3A)
# pdf('../plots/figure_mobility/fit_gam_safegraph_counties.pdf',
# height = 3., width = 3.9)
plot(plt_fit_mobile_phone_counties_with_R2 +
theme(legend.position = c(0.35, 0.8),
legend.background = element_blank()) +
guides(colour = guide_legend(override.aes = list(alpha = 1.))))
#dev.off()
## Outliers from GAM between RR of id seq and RR of movement from mobile phone data (Figure 3B)
#pdf('../plots/figure_mobility/outliers_gam_safegraph_counties.pdf',
# height = 3., width = 3.9)
plot(plt_outliers_mobile_phone_mobility_counties + theme(legend.position = 'none'))
#dev.off()
## Combination of Figure 3A and 3B (with the same legend)
# pdf('../plots/figure_mobility/panel_gam_safegraph_counties_with_R2.pdf',
# height = 3.5, width = 8.)
plot(ggarrange(plt_fit_mobile_phone_counties_with_R2 +
guides(colour = guide_legend(override.aes = list(alpha = 1.))),
plt_outliers_mobile_phone_mobility_counties, common.legend = T, legend = 'top'))
#dev.off()
## Figure with the results of the fit of the RR of id sequences to the 3 mobility indicators at the county level (Figure S10)
panel_fit_counties <- ggarrange(plt_fit_mobile_phone_counties_with_R2 +
theme(legend.background = element_blank(), legend.position = c(0.3, 0.8)) +
guides(colour = guide_legend(override.aes = list(alpha = 1.)),
alpha = 'none') +
ggtitle('Mobile phone mobility data'),
plt_outliers_mobile_phone_mobility_counties +
theme(legend.position = 'none') +
ggtitle('Mobile phone mobility data'),
plt_fit_workflow_counties_with_R2 +
theme(legend.position = 'none') +
ggtitle('Commuting mobility data'),
plt_outliers_workflow_mobility_counties +
theme(legend.background = element_blank(), legend.position = c(0.7, 0.92)) +
guides(colour = guide_legend(override.aes = list(alpha = 1.)),
alpha = 'none') +
ggtitle('Commuting mobility data'),
plt_fit_distance_counties_with_R2 +
theme(legend.position = 'none') +
ggtitle('Distance'),
plt_outliers_distance_mobility_counties +
theme(legend.background = element_blank(), legend.position = c(0.7, 0.95)) +
guides(colour = guide_legend(override.aes = list(alpha = 1.)),
alpha = 'none') +
ggtitle('Distance'),
labels = 'AUTO', nrow = 3, ncol = 2)
# pdf('../plots/figure_mobility/summary_gam_safegraph_workflow_counties.pdf',
# height = 10.5, width = 9.0)
plot(panel_fit_counties)
# dev.off()
## Figure with the results of the fit of the RR of id sequences to the 3 mobility indicators at the regional level (Figure S12)
panel_fit_regions <- ggarrange(plt_fit_mobile_phone_regions_with_R2 + ggtitle('Mobile phone mobility data'),
plt_fit_workflow_regions_with_R2 + ggtitle('Commuting mobility data'),
plt_fit_distance_regions_with_R2 + ggtitle('Distance'),
plt_residuals_regions, labels = 'AUTO')
# pdf('../plots/figure_mobility/summary_gam_safegraph_workflow_regions.pdf',
# height = 6.5, width = 7.5)
plot(panel_fit_regions)
#dev.off()
## Look at a couple values of the RR of identical sequences between counties
df_RR_counties %>% filter(group_1 == 'Franklin County', group_2 == 'Mason County')
df_RR_counties %>% filter(group_1 == 'Walla Walla County', group_2 == 'Mason County')
df_RR_counties %>% filter(group_1 == 'Pierce County', group_2 == 'Mason County')
df_RR_uncertainty_counties %>% filter(group_1 == 'Franklin County', group_2 == 'Mason County')
df_RR_uncertainty_counties %>% filter(group_1 == 'Walla Walla County', group_2 == 'Mason County')
df_RR_uncertainty_counties %>% filter(group_1 == 'Pierce County', group_2 == 'Mason County')
## Look at outliers in the relationship between genetic and mobility data (for our 3 mobility indicators)
df_outliers_workflow_mobility %>%
mutate(type = 'Workflow') %>%
bind_rows(df_outliers_mobile_phone_mobility %>%
mutate(type = 'Mobile phone')) %>%
bind_rows(df_outliers_distance_mobility %>%
mutate(type = 'Distance'))