Liza Brusman 2024-10-29
library(dplyr)
library(tidyr)
library(ggplot2)
library(ggridges)
library(forcats)
library(ggpubr)
library(glmmTMB)
library(DHARMa)
library(gridExtra)
library(emmeans)
library(Hmisc)
library(corrplot)
source("SVM_plot_fxns.R")metadata
metadata <- read.csv("../../docs/seq_beh_metadata.csv")
metadata$Group <- paste(metadata$sex, metadata$SS_OS, sep = "_")
metadata <- metadata %>% filter(!animal %in% c("4918", "4967"))module data
all_cells <- read.csv("../hotspot/docs/ani_mod_scores_allcells_lognorm_counts.csv")
modules <- colnames(all_cells)
modules <- modules[modules != "animal"]
data <- merge(all_cells, metadata, on = "animal")for SVMs that include self:
with_self <- read.csv("output/ani_pairwise_classifications_withself.csv")for SVMs that do not include self
exclude_self <- read.csv("output/all_svm_combined_excludeself_200cells.csv") %>% filter(animal == Var1)put clusters in right order and assign colors
f.levels <- c('Drd1Pdyn', 'Drd1PdynOprm1', 'Drd1Penk', 'Drd2Penk', 'Drd2NoPenk',
'GABAergicNeurons', 'Dlx2ImmatureNeurons', 'SstNpyInterneurons',
'PvalbInterneurons', 'CholinergicInterneurons', 'MatureOligos',
'ImmatureOligos', 'Astrocytes', 'Microglia')
clust_cols <- c("Drd1Pdyn" = "#304880", "Drd1PdynOprm1" = "#4898D0", "Drd1Penk" = "#88C0F0", "Drd2Penk" = "#9060A0", "Drd2NoPenk" = "#C078C0", "GABAergicNeurons" = "#389078", "Dlx2ImmatureNeurons" = "#70A830", "SstNpyInterneurons" = "#98D048", "PvalbInterneurons" = "#60D0A0", "CholinergicInterneurons" = "#80E8C0", "MatureOligos" = "#783028", "ImmatureOligos" = "#B82820", "Astrocytes" = "#D04058", "Microglia" = "#F87878")process df into format for ridgeline plots
with_self2 <- process_df_withself(with_self)## `summarise()` has grouped output by 'Cluster'. You can override using the
## `.groups` argument.
plot ridgeline plot
plot_ridgeline(with_self2, save = FALSE)## Picking joint bandwidth of 0.0453
STAT TESTS - GLM for with self - self vs. partner vs. other
with_self_stats <- GLM_SVM(with_self2)
# setwd("output/")
# write.csv(with_self_stats, "glm_svm_with_self_groups_ordbeta.csv")widen dataframe for other analyses later
with_self_wide <- widen_df_withself(with_self2)process df for excluding self
exclude_self2 <- process_df_excludeself(exclude_self)## `summarise()` has grouped output by 'Cluster'. You can override using the
## `.groups` argument.
plot exclude self ridgeline
plot_ridgeline(exclude_self2, save = FALSE)## Picking joint bandwidth of 0.0526
GLM stats: exclude self
exclude_self_stats <- GLM_SVM(exclude_self2)
# setwd("output/")
# write.csv(exclude_self_stats, "svm_exclude_self_glm_groups.csv")exclude self widened df
exclude_self_wide <- widen_df_excludeself(exclude_self2)plot ridgelines by pairing type - with self
pair_types <- unique(data$pair_type)
for (type in pair_types) {
print(type)
plot_ridgeline(with_self2, grouping = "pair_type", save = FALSE)
}## [1] "FM"
## Picking joint bandwidth of 0.0477
## [1] "MM"
## Picking joint bandwidth of 0.0549
## [1] "FF"
## Picking joint bandwidth of 0.0623
ridgeline by pair type without self
pair_types <- unique(data$pair_type)
for (type in pair_types) {
print(type)
plot_ridgeline(exclude_self2, grouping = "pair_type", save = FALSE)
}## [1] "FM"
## Picking joint bandwidth of 0.0527
## [1] "MM"
## Picking joint bandwidth of 0.0699
## [1] "FF"
## Picking joint bandwidth of 0.0608
ridgeline by cohort with self
cohs <- unique(data$beh_cohort)
for (type in cohs) {
print(type)
plot_ridgeline(with_self2, grouping = "beh_cohort", save = FALSE)
}## [1] 3
## Picking joint bandwidth of 0.0539
## [1] 2
## Picking joint bandwidth of 0.0565
## [1] 4
## Picking joint bandwidth of 0.0504
ridgelines by cohort exclude self
cohs <- unique(data$beh_cohort)
for (type in cohs) {
print(type)
plot_ridgeline(exclude_self2, grouping = "beh_cohort", save = FALSE)
}## [1] 3
## Picking joint bandwidth of 0.0698
## [1] 2
## Picking joint bandwidth of 0.0637
## [1] 4
## Picking joint bandwidth of 0.0522
directly compare pairing types. each point in these plots is the percentage of classification for one animal to one other animal
exclude_self2$group_type <- paste(exclude_self2$pair_type, exclude_self2$group, sep = "_")
plot_dotcloud(exclude_self2)
for stats - looking at partner/other*pair_type
exclude_self_pairtype_stats <- GLM_SVM_pairtype(exclude_self2)
# setwd("output/")
# write.csv(exclude_self_pairtype_stats, "svm_excludeself_glm_posthocs_by_pairtype.csv")correlation between partners in partner % classification
plot_SVM_corrs(exclude_self_wide, save = FALSE)## `geom_smooth()` using formula = 'y ~ x'
## `geom_smooth()` using formula = 'y ~ x'
## `geom_smooth()` using formula = 'y ~ x'
## `geom_smooth()` using formula = 'y ~ x'
## `geom_smooth()` using formula = 'y ~ x'
## `geom_smooth()` using formula = 'y ~ x'
## `geom_smooth()` using formula = 'y ~ x'
## `geom_smooth()` using formula = 'y ~ x'
## `geom_smooth()` using formula = 'y ~ x'
## `geom_smooth()` using formula = 'y ~ x'
## `geom_smooth()` using formula = 'y ~ x'
## `geom_smooth()` using formula = 'y ~ x'
## `geom_smooth()` using formula = 'y ~ x'
## `geom_smooth()` using formula = 'y ~ x'
## `geom_smooth()` using formula = 'y ~ x'
## `geom_smooth()` using formula = 'y ~ x'
## `geom_smooth()` using formula = 'y ~ x'
## `geom_smooth()` using formula = 'y ~ x'
## `geom_smooth()` using formula = 'y ~ x'
## `geom_smooth()` using formula = 'y ~ x'
## `geom_smooth()` using formula = 'y ~ x'
## `geom_smooth()` using formula = 'y ~ x'
## `geom_smooth()` using formula = 'y ~ x'
## `geom_smooth()` using formula = 'y ~ x'
## `geom_smooth()` using formula = 'y ~ x'
## `geom_smooth()` using formula = 'y ~ x'
## `geom_smooth()` using formula = 'y ~ x'
## `geom_smooth()` using formula = 'y ~ x'
get out some summary stats for exclude self data
summ_stats <- exclude_self_wide %>% group_by(Cluster) %>% summarise(mean_partner_pct = mean(partner.pct), avg_cells_tested = mean(partner+other))
ggplot(summ_stats, aes(x = avg_cells_tested, y = mean_partner_pct, color = Cluster)) +
geom_point() +
geom_smooth(method = "lm") +
scale_color_manual(values = clust_cols)## `geom_smooth()` using formula = 'y ~ x'
#here i correlated the mean partner % vs. the number of cells tested
rcorr(summ_stats$mean_partner_pct, summ_stats$avg_cells_tested)## x y
## x 1.00 0.49
## y 0.49 1.00
##
## n= 14
##
##
## P
## x y
## x 0.0727
## y 0.0727
get some summary stats for with self data
summ_stats <- with_self_wide %>% group_by(Cluster) %>% summarise(mean_self_pct = mean(self.pct), avg_cells_tested = mean(self+partner+other))
#here i correlated the self classifications to the number of cells tested. this is highly correlated
cor_test <- rcorr(summ_stats$mean_self_pct, summ_stats$avg_cells_tested, type = "spearman")
print(cor_test)## x y
## x 1.00 0.88
## y 0.88 1.00
##
## n
## x y
## x 13 13
## y 13 14
##
## P
## x y
## x 0
## y 0
ggplot(summ_stats, aes(x = avg_cells_tested, y = mean_self_pct, color = Cluster)) +
geom_point() +
geom_smooth(method = "lm") +
scale_color_manual(values = clust_cols)## `geom_smooth()` using formula = 'y ~ x'
self vs. chance - one-sample t-test vs. 1/38 (correct classification by random)
vs_chance_df <- data.frame()
for (clust in f.levels) {
clust_df <- with_self2 %>% filter(Cluster == clust) %>% filter(group == "self")
onesamp <- t.test(clust_df$norm_classes, mu = 1/38)
# print(oneway)
stats_df <- data.frame(Cluster = clust,
T_val = onesamp$statistic[1],
p.value = onesamp$p.value,
test = "onesamp t-test")
vs_chance_df <- rbind(vs_chance_df, stats_df)
}partner vs. chance for exclude self
vs_chance_df_exclude <- data.frame()
for (clust in f.levels) {
clust_df <- exclude_self2 %>% filter(Cluster == clust) %>% filter(group == "partner")
onesamp <- t.test(clust_df$norm_classes, mu = 1/37)
# print(oneway)
stats_df <- data.frame(Cluster = clust,
T_val = onesamp$statistic[1],
p.value = onesamp$p.value,
test = "onesamp t-test")
vs_chance_df_exclude <- rbind(vs_chance_df_exclude, stats_df)
}look at classifications for same sex, same cohort, same pairing type, relatedness, etc. basically - what other factors are at play other than partner ID?
exclude_self_newclass <- exclude_self2
exclude_self_newclass <- exclude_self_newclass %>% merge(metadata, on = "animal")
#basically just creating named lists to match to predicted animal
sex_vec <- metadata$sex
names(sex_vec) <- metadata$animal
exclude_self_newclass$y_pred <- as.character(exclude_self_newclass$y_pred)
exclude_self_newclass$same_opp_sex <- if_else(exclude_self_newclass$sex == sex_vec[exclude_self_newclass$y_pred], "same", "different")
coh_vec <- metadata$beh_cohort
names(coh_vec) <- metadata$animal
exclude_self_newclass$same_diff_coh <- if_else(exclude_self_newclass$beh_cohort == coh_vec[exclude_self_newclass$y_pred], "same", "different")
type_vec <- metadata$pair_type
names(type_vec) <- metadata$animal
exclude_self_newclass$same_diff_type <- if_else(exclude_self_newclass$pair_type == type_vec[exclude_self_newclass$y_pred], "same", "different")
group_vec <- metadata$Group
names(group_vec) <- metadata$animal
exclude_self_newclass$same_diff_group <- if_else(exclude_self_newclass$Group == group_vec[exclude_self_newclass$y_pred], "same", "different")
parent_vec <- metadata$parents
names(parent_vec) <- metadata$animal
exclude_self_newclass$same_diff_parents <- if_else(exclude_self_newclass$parents == parent_vec[exclude_self_newclass$y_pred], "same", "different")only look at “other” classifications
exclude_self_newclass_other <- exclude_self_newclass %>% filter(group == "other")
norm_to_other_class <- exclude_self_newclass_other %>% group_by(Cluster, animal) %>% summarise(sum_other_classes = sum(Freq))## `summarise()` has grouped output by 'Cluster'. You can override using the
## `.groups` argument.
same_sex_norm <- exclude_self_newclass_other %>% group_by(Cluster, animal, same_opp_sex) %>% summarise(classes = sum(Freq))## `summarise()` has grouped output by 'Cluster', 'animal'. You can override using
## the `.groups` argument.
same_sex_norm$metric <- "sex"
same_sex_norm <- same_sex_norm %>% rename("same_diff" = "same_opp_sex")
same_sex_norm <- same_sex_norm %>% merge(norm_to_other_class, on = "animal")
same_coh_norm <- exclude_self_newclass_other %>% group_by(Cluster, animal, same_diff_coh) %>% summarise(classes = sum(Freq))## `summarise()` has grouped output by 'Cluster', 'animal'. You can override using
## the `.groups` argument.
same_coh_norm$metric <- "cohort"
same_coh_norm <- same_coh_norm %>% rename("same_diff" = "same_diff_coh")
same_coh_norm <- same_coh_norm %>% merge(norm_to_other_class, on = "animal")
same_type_norm <- exclude_self_newclass_other %>% group_by(Cluster, animal, same_diff_type) %>% summarise(classes = sum(Freq))## `summarise()` has grouped output by 'Cluster', 'animal'. You can override using
## the `.groups` argument.
same_type_norm$metric <- "pair_type"
same_type_norm <- same_type_norm %>% rename("same_diff" = "same_diff_type")
same_type_norm <- same_type_norm %>% merge(norm_to_other_class, on = "animal")
same_group_norm <- exclude_self_newclass_other %>% group_by(Cluster, animal, same_diff_group) %>% summarise(classes = sum(Freq))## `summarise()` has grouped output by 'Cluster', 'animal'. You can override using
## the `.groups` argument.
same_group_norm$metric <- "group"
same_group_norm <- same_group_norm %>% rename("same_diff" = "same_diff_group")
same_group_norm <- same_group_norm %>% merge(norm_to_other_class, on = "animal")
same_parents_norm <- exclude_self_newclass_other %>% group_by(Cluster, animal, same_diff_parents) %>% summarise(classes = sum(Freq))## `summarise()` has grouped output by 'Cluster', 'animal'. You can override using
## the `.groups` argument.
same_parents_norm$metric <- "parents"
same_parents_norm <- same_parents_norm %>% rename("same_diff" = "same_diff_parents")
same_parents_norm <- same_parents_norm %>% merge(norm_to_other_class, on = "animal")
# norm_to_other_class <- norm_to_other_class %>% merge(same_sex_norm, on = "animal") %>% merge(same_coh_norm, on = "animal") %>% merge(same_type_norm, on = "animal") %>% merge(same_group_norm, on = "animal") %>% merge(same_parents_norm, on = "animal")
norm_to_other_class <- rbind(same_sex_norm, same_coh_norm, same_type_norm, same_group_norm, same_parents_norm)
norm_to_other_class$pcts <- norm_to_other_class$classes/norm_to_other_class$sum_other_classesnormalize to the percent of classifications that are for “other” (i.e. what percent of “other” classifications?)
norm_to_other_pct <- exclude_self_newclass_other %>% group_by(Cluster, animal) %>% summarise(sum_other_classes = sum(Freq))## `summarise()` has grouped output by 'Cluster'. You can override using the
## `.groups` argument.
exclude_self_newclass_other <- merge(exclude_self_newclass_other, norm_to_other_pct, on = "animal")
exclude_self_newclass_other$norm_to_other <- exclude_self_newclass_other$Freq/exclude_self_newclass_other$sum_other_classesplot misclassification allocations by cluster
clust_cols <- c("Drd1Pdyn" = "#304880",
"Drd1PdynOprm1" = "#4898D0",
"Drd1Penk" = "#88C0F0",
"Drd2Penk" = "#9060A0",
"Drd2NoPenk" = "#C078C0",
"GABAergicNeurons" = "#389078",
"Dlx2ImmatureNeurons" = "#70A830",
"SstNpyInterneurons" = "#98D048",
"PvalbInterneurons" = "#60D0A0",
"CholinergicInterneurons" = "#80E8C0",
"MatureOligos" = "#783028",
"ImmatureOligos" = "#B82820",
"Astrocytes" = "#D04058",
"Microglia" = "#F87878")
clust_cols_new <- clust_cols
names(clust_cols_new) <- paste0(names(clust_cols), "_same")
clust_cols_light <- colorspace::lighten(clust_cols, 0.6)
clust_cols_light <- rep("gray69", 14)
names(clust_cols_light) <- paste0(names(clust_cols), "_different")
idx <- order(c(seq_along(clust_cols_new), seq_along(clust_cols_light)))
all_clust_cols <- (c(clust_cols_new,clust_cols_light))[idx]
newclass_long_pcts <- exclude_self_newclass_other %>% pivot_longer(cols = c("same_opp_sex", "same_diff_coh", "same_diff_type", "same_diff_group", "same_diff_parents"))
newclass_long_pcts$value <- factor(newclass_long_pcts$value, levels = c("same", "different"))
newclass_long_pcts$Cluster <- factor(newclass_long_pcts$Cluster, levels = f.levels)
newclass_long_pcts$Cluster_samediff <- paste0(newclass_long_pcts$Cluster, "_", newclass_long_pcts$value)
p <- ggplot(newclass_long_pcts, aes(x = factor(Cluster_samediff, levels = names(all_clust_cols)), y = norm_to_other*100, color = Cluster_samediff, fill = Cluster_samediff, alpha = 0.8)) +
# geom_boxplot() +
# geom_point(position = position_jitterdodge(jitter.width = 0, dodge.width = 0.9)) +
geom_jitter(size = 0.5, width = 0.2) +
scale_color_manual(values = all_clust_cols) +
scale_fill_manual(values = all_clust_cols) +
xlab("Group") +
ylab("% of Classifications") +
theme_classic()
p2 <- p + facet_wrap(~ name, nrow = 5)
print(p2)
# setwd("output/new_plots/")
# ggsave("by_metrics_animal_by_animal.pdf", p2)plot allocations of misclassifications as barplot
plt_list <- list()
for (m in c("same_diff_coh", "same_diff_group", "same_diff_type", "same_diff_parents", "same_opp_sex")) {
print(m)
#first find expected values
metadata2 <- metadata
if (m == "same_opp_sex") {
metadata2 <- metadata2 %>% rename(m = "sex")
}
else if (m == "same_diff_group") {
metadata2 <- metadata2 %>% rename(m = "Group")
}
else if(m == "same_diff_type") {
metadata2 <- metadata2 %>% rename(m = "pair_type")
}
else if(m == "same_diff_parents") {
metadata2 <- metadata2 %>% rename(m = "parents")
}
else if(m == "same_diff_coh") {
metadata2 <- metadata2 %>% rename(m = "beh_cohort")
}
#get weighted avg expected value
m_summ <- metadata2 %>% summarise(n_metric = n(), .by = m)
m_summ$norm_pct <- m_summ$n_metric/38
weighted_mean <- weighted.mean(m_summ$norm_pct, m_summ$n_metric)
print(weighted_mean)
#make df to sum all other classifications, don't group by cluster first
test_m <- exclude_self_newclass_other %>% group_by_at(c("animal", "group", m)) %>% summarise_at(.vars = "Freq", .funs = sum)
test_m <- test_m %>% rename("same_diff" = m)
test_m2 <- test_m %>% group_by_at(c("animal", "group")) %>% summarise(total_other_classifications = sum(Freq)) %>% merge(test_m, on = "animal") %>% filter(same_diff == "same")
test_m2$pct_same_classes <- test_m2$Freq/test_m2$total_other_classifications
test_m2$metric <- m
#one sample ttest vs. chance line
one_samp_ttest <- t.test(test_m2$pct_same_classes*100, mu = weighted_mean*100)
print(one_samp_ttest)
p3 <- ggplot(test_m2, aes(x = metric, y = pct_same_classes*100)) +
geom_bar(stat = "summary", fun = "mean", color = "black", fill = "white") +
stat_summary(
fun.data = mean_se, fun.args = list(mult = 1), geom = "errorbar", color = "black", #fun.args = list(mult = 1),
position = position_dodge(0.8), width = 0.2, linewidth = 0.8
) +
geom_jitter(height=0, width = 0.2, color = "black", alpha = 0.6, stroke = NA) +
geom_hline(yintercept = weighted_mean*100, color = "red", linetype = "dashed") +
ylim(0, 100) +
theme_classic()
print(p3)
plt_list[[m]] <- p3
}## [1] "same_diff_coh"
## [1] 0.3462604
## `summarise()` has grouped output by 'animal'. You can override using the
## `.groups` argument.
##
## One Sample t-test
##
## data: test_m2$pct_same_classes * 100
## t = 45.411, df = 37, p-value < 2.2e-16
## alternative hypothesis: true mean is not equal to 34.62604
## 95 percent confidence interval:
## 84.55940 89.22346
## sample estimates:
## mean of x
## 86.89143
## [1] "same_diff_group"
## [1] 0.2520776
## `summarise()` has grouped output by 'animal'. You can override using the
## `.groups` argument.
##
## One Sample t-test
##
## data: test_m2$pct_same_classes * 100
## t = -7.2743, df = 37, p-value = 1.226e-08
## alternative hypothesis: true mean is not equal to 25.20776
## 95 percent confidence interval:
## 7.989812 15.491970
## sample estimates:
## mean of x
## 11.74089
## [1] "same_diff_type"
## [1] 0.3905817
## `summarise()` has grouped output by 'animal'. You can override using the
## `.groups` argument.
##
## One Sample t-test
##
## data: test_m2$pct_same_classes * 100
## t = -4.6812, df = 37, p-value = 3.76e-05
## alternative hypothesis: true mean is not equal to 39.05817
## 95 percent confidence interval:
## 13.11857 28.79036
## sample estimates:
## mean of x
## 20.95446
## [1] "same_diff_parents"
## [1] 0.09695291
## `summarise()` has grouped output by 'animal'. You can override using the
## `.groups` argument.
##
## One Sample t-test
##
## data: test_m2$pct_same_classes * 100
## t = 2.6022, df = 35, p-value = 0.01349
## alternative hypothesis: true mean is not equal to 9.695291
## 95 percent confidence interval:
## 11.55082 24.72063
## sample estimates:
## mean of x
## 18.13572
## [1] "same_opp_sex"
## [1] 0.501385
## `summarise()` has grouped output by 'animal'. You can override using the
## `.groups` argument.
##
## One Sample t-test
##
## data: test_m2$pct_same_classes * 100
## t = -1.526, df = 37, p-value = 0.1355
## alternative hypothesis: true mean is not equal to 50.1385
## 95 percent confidence interval:
## 37.54385 51.91185
## sample estimates:
## mean of x
## 44.72785
p4 <- grid.arrange(grobs = plt_list, nrow = 1)
print(p4)## TableGrob (1 x 5) "arrange": 5 grobs
## z cells name grob
## same_diff_coh 1 (1-1,1-1) arrange gtable[layout]
## same_diff_group 2 (1-1,2-2) arrange gtable[layout]
## same_diff_type 3 (1-1,3-3) arrange gtable[layout]
## same_diff_parents 4 (1-1,4-4) arrange gtable[layout]
## same_opp_sex 5 (1-1,5-5) arrange gtable[layout]
setwd("output/new_plots/")
# ggsave("by_metrics_percent_cells_no_partner_combine_clusts_bars_se_noclustcollapsefirst.pdf", p4, width = 7, height = 7)try a way to visualize only means (from data normalized to “other” counts) this sets up dataframe to make a bubble plot
norm_to_other_class2 <- norm_to_other_class %>% filter(same_diff == "same")
norm_for_summary <- norm_to_other_class2 %>% group_by(Cluster, metric) %>% summarise(mean_same_pct = mean(pcts, na.rm = TRUE))## `summarise()` has grouped output by 'Cluster'. You can override using the
## `.groups` argument.
bubble plot with normalized avg expected values colored bubble indicates true mean black line indicates expected (weighted mean)
norm_wts <- data.frame()
for (m in unique(norm_for_summary$metric)) {
print(m)
m_df <- norm_for_summary %>% filter(metric == m)
metadata2 <- metadata %>% filter(!animal %in% c("4918", "4967"))
if (m == "cohort") {
metadata2 <- metadata2 %>% rename(m = "beh_cohort")
}
else if (m == "group") {
metadata2 <- metadata2 %>% rename(m = "Group")
}
#get weighted avg expected value
m_summ <- metadata2 %>% summarise(n_metric = n(), .by = m)
m_summ$norm_pct <- m_summ$n_metric/38
weighted_mean <- weighted.mean(m_summ$norm_pct, m_summ$n_metric)
mini_df <- data.frame("metric" = m, "weighted_mean" = weighted_mean)
norm_wts <- rbind(norm_wts, mini_df)
}## [1] "cohort"
## [1] "group"
## [1] "pair_type"
## [1] "parents"
## [1] "sex"
norm_for_summary_weighted <- norm_for_summary %>% merge(norm_wts, on = "metric")
p <- ggplot(norm_for_summary_weighted, aes(x = metric, y = factor(Cluster, levels = rev(f.levels)), fill = Cluster, color = Cluster)) +
geom_point(aes(size = mean_same_pct), alpha = 1, shape = 21) +
geom_point(aes(size = weighted_mean), alpha = 1, shape = 1, color = "black") +
# geom_point(aes(size = 1), alpha = 1, shape = 1, color = "black") + #if you want 100% circle
scale_size_continuous(limits = c(0, 1), range = c(0, 10)) +
scale_fill_manual(values = clust_cols, guide = FALSE) +
scale_color_manual(values = clust_cols, guide = FALSE) +
theme(panel.background = element_blank(), panel.border = element_rect(colour = "black", fill = NA, size = 1.2))
print(p)
# setwd("output/new_plots/")
# ggsave("SVM_metric_classifications_bubbleplot_normalized_expected.pdf", p)sessionInfo()## R version 4.2.2 (2022-10-31 ucrt)
## Platform: x86_64-w64-mingw32/x64 (64-bit)
## Running under: Windows 10 x64 (build 22631)
##
## Matrix products: default
##
## locale:
## [1] LC_COLLATE=English_United States.utf8
## [2] LC_CTYPE=English_United States.utf8
## [3] LC_MONETARY=English_United States.utf8
## [4] LC_NUMERIC=C
## [5] LC_TIME=English_United States.utf8
##
## attached base packages:
## [1] stats graphics grDevices utils datasets methods base
##
## other attached packages:
## [1] corrplot_0.92 Hmisc_5.0-1 emmeans_1.8.5 gridExtra_2.3 DHARMa_0.4.6
## [6] glmmTMB_1.1.8 ggpubr_0.6.0 forcats_1.0.0 ggridges_0.5.4 ggplot2_3.4.2
## [11] tidyr_1.3.0 dplyr_1.1.1
##
## loaded via a namespace (and not attached):
## [1] splines_4.2.2 carData_3.0-5 Formula_1.2-5
## [4] highr_0.10 yaml_2.3.7 numDeriv_2016.8-1.1
## [7] pillar_1.9.0 backports_1.4.1 lattice_0.21-8
## [10] glue_1.6.2 digest_0.6.31 ggsignif_0.6.4
## [13] checkmate_2.1.0 minqa_1.2.5 colorspace_2.1-0
## [16] sandwich_3.0-2 htmltools_0.5.5 Matrix_1.6-2
## [19] pkgconfig_2.0.3 broom_1.0.4 purrr_1.0.1
## [22] xtable_1.8-4 mvtnorm_1.1-3 scales_1.2.1
## [25] lme4_1.1-33 tibble_3.2.1 htmlTable_2.4.1
## [28] mgcv_1.8-42 farver_2.1.1 generics_0.1.3
## [31] car_3.1-2 TH.data_1.1-2 withr_2.5.0
## [34] nnet_7.3-18 TMB_1.9.10 cli_3.6.0
## [37] survival_3.5-5 magrittr_2.0.3 estimability_1.4.1
## [40] evaluate_0.20 fansi_1.0.4 nlme_3.1-162
## [43] MASS_7.3-58.3 rstatix_0.7.2 foreign_0.8-84
## [46] data.table_1.14.6 tools_4.2.2 lifecycle_1.0.3
## [49] multcomp_1.4-23 stringr_1.5.0 munsell_0.5.0
## [52] cluster_2.1.4 compiler_4.2.2 rlang_1.1.1
## [55] grid_4.2.2 nloptr_2.0.3 rstudioapi_0.14
## [58] htmlwidgets_1.6.2 labeling_0.4.2 base64enc_0.1-3
## [61] rmarkdown_2.25 boot_1.3-28.1 gtable_0.3.3
## [64] codetools_0.2-19 abind_1.4-5 R6_2.5.1
## [67] zoo_1.8-12 knitr_1.42 fastmap_1.1.1
## [70] utf8_1.2.3 stringi_1.7.12 Rcpp_1.0.10
## [73] vctrs_0.6.1 rpart_4.1.19 tidyselect_1.2.0
## [76] xfun_0.38 coda_0.19-4
Add a new chunk by clicking the Insert Chunk button on the toolbar or by pressing Ctrl+Alt+I.
When you save the notebook, an HTML file containing the code and output will be saved alongside it (click the Preview button or press Ctrl+Shift+K to preview the HTML file).
The preview shows you a rendered HTML copy of the contents of the editor. Consequently, unlike Knit, Preview does not run any R code chunks. Instead, the output of the chunk when it was last run in the editor is displayed.