overlaps0[i] <- if_else(c(rmats_event_th_pi_mod2_hyp_df$CI.Lower[i], rmats_event_th_pi_mod2_hyp_df$CI.Upper[i]) %overlaps% c(0,0), "black",mypal[1])
}
rmats_event_th_pi_mod2_hyp_df$color <- overlaps0}
rmats_event_th_pi_mod2_hyp_plot <- hyp_plot +
geom_point(data = rmats_event_th_pi_mod2_hyp_df, aes(x = Hypothesis, y = Estimate, color = color)) +
geom_errorbar(data = rmats_event_th_pi_mod2_hyp_df, aes(x = Hypothesis, color = color, ymin = CI.Lower, ymax = CI.Upper), width = 0.1)
rbind(conditional_effects(rmats_event_ab_omega_mod2)[[1]][,5:9] %>%  mutate(response = "omega", tissue = "Abdomen"),
conditional_effects(rmats_event_ab_pi_mod2)[[1]][,4:8] %>%  mutate(response = "pi", tissue = "Abdomen"),
conditional_effects(rmats_event_th_omega_mod2)[[1]][,5:9] %>%  mutate(response = "omega", tissue = "Thorax"),
conditional_effects(rmats_event_th_pi_mod2)[[1]][,4:8] %>%  mutate(response = "pi", tissue = "Thorax")
) %>%
ggplot(aes(x = effect1__, y = estimate__)) +
geom_errorbar(aes(ymin = lower__, ymax = upper__), width = 0.25) +
geom_point() +
labs( x = NULL, y ="Estimate")+
theme_bw() +
facet_grid(response~ tissue, scales = "free")
ggarrange(NULL, NULL,
rmats_event_ab_pi_mod2_hyp_plot,  rmats_event_ab_omega_mod2_hyp_plot,
rmats_event_th_pi_mod2_hyp_plot,  rmats_event_th_omega_mod2_hyp_plot,
ncol = 2,
nrow = 3, labels = c(NA, NA, "Abdomen pi", "Abdomen omega", "Thorax pi", "Thorax omega"), heights = c(0.1, 1,1), vjust = 0)
mypal<- ggsci::pal_nejm()(8)
mypal3<- adjustcolor(c("black",  mypal), alpha.f = 0.75)
events <- c("A3SS", "A5SS", "SE", "IR", "MXE")
# save(rMATS_filt20_AB_DRY_vs_AB_WET, file = "rMATS_filt20_AB_DRY_vs_AB_WET.Rdata")
load("/Users/rachelsteward/OneDrive/B_anynana_altSplicing/B_anynana_diff_expr_analysis/B_anynana_rMATS/rMATS_filt3ind_AB_DRY_vs_AB_WET.Rdata")
rMATS_filt3ind_AB_DRY_vs_AB_WET <- rMATS_filt3
# save(rMATS_filt20_TH_DRY_vs_TH_WET, file = "rMATS_filt20_TH_DRY_vs_TH_WET.Rdata")
load("/Users/rachelsteward/OneDrive/B_anynana_altSplicing/B_anynana_diff_expr_analysis/B_anynana_rMATS/rMATS_filt3ind_TH_DRY_vs_TH_WET.Rdata")
rMATS_filt3ind_TH_DRY_vs_TH_WET <- rMATS_filt3
rm(rMATS_filt3)
#### Abdomen DF ####
rmats_ab <- data.frame()
for (i in 1:length(rMATS_filt3ind_AB_DRY_vs_AB_WET)) {
temp <- rMATS_filt3ind_AB_DRY_vs_AB_WET[[i]] %>% select(ID, geneSymbol, IncFormLen, SkipFormLen, FDR, IncLevelDifference, event, comp)
rmats_ab <- rbind(rmats_ab, temp)
}
rmats_ab$event <- factor(rmats_ab$event, levels = events)
rmats_ab <- rmats_ab %>% rename(GeneID=geneSymbol)
rmats_ab_splice_summary <- rmats_ab %>% group_by(event) %>%
summarize(n_spl_events = n(),
genes = length(unique(GeneID)),
spl_event_per_gene = n_spl_events/genes,
n_sig_spl_events = sum(as.numeric(FDR) < 0.05 & abs(IncLevelDifference) > 0.05)) %>% ungroup %>%
mutate(total_spl_events = sum(n_spl_events) ,
total_sig_spl_events = sum(n_sig_spl_events),
expected_spl_events = (n_spl_events/total_spl_events)*total_sig_spl_events,
event = factor(event, levels = c("A3SS", "A5SS", "SE", "IR", "MXE")))
# load DS_ab_allgenes from nucleotideUpset.R
rmats_ab_summ <- rmats_ab %>%  group_by(GeneID,event) %>%
summarize(all_events = n(),
all_sig_events = sum(FDR < 0.05 & abs(IncLevelDifference) > 0.05, na.rm = T),
sig_DS_event = if_else(all_sig_events > 0, "DS", "AS")) %>%
full_join(data.frame(unique(na.omit(DS_ab_allgenes))))%>%
mutate(tissue = "Abdomen")
## DS abdomen
DS_ab_allgenes <- read_tsv("/Users/rachelsteward/OneDrive/B_anynana_altSplicing/PanicTime/edgeR_DS_simesDataframe_family_abdomen.tsv") %>% select(GeneID) %>% mutate(GeneID = gsub("gene-", "", GeneID))
## DS thorax
DS_th_allgenes <- read_tsv("/Users/rachelsteward/OneDrive/B_anynana_altSplicing/PanicTime/edgeR_DS_simesDataframe_family_thorax.tsv") %>% select(GeneID) %>% mutate(GeneID = gsub("gene-", "", GeneID))
mypal<- ggsci::pal_nejm()(8)
mypal3<- adjustcolor(c("black",  mypal), alpha.f = 0.75)
events <- c("A3SS", "A5SS", "SE", "IR", "MXE")
# save(rMATS_filt20_AB_DRY_vs_AB_WET, file = "rMATS_filt20_AB_DRY_vs_AB_WET.Rdata")
load("/Users/rachelsteward/OneDrive/B_anynana_altSplicing/B_anynana_diff_expr_analysis/B_anynana_rMATS/rMATS_filt3ind_AB_DRY_vs_AB_WET.Rdata")
rMATS_filt3ind_AB_DRY_vs_AB_WET <- rMATS_filt3
# save(rMATS_filt20_TH_DRY_vs_TH_WET, file = "rMATS_filt20_TH_DRY_vs_TH_WET.Rdata")
load("/Users/rachelsteward/OneDrive/B_anynana_altSplicing/B_anynana_diff_expr_analysis/B_anynana_rMATS/rMATS_filt3ind_TH_DRY_vs_TH_WET.Rdata")
rMATS_filt3ind_TH_DRY_vs_TH_WET <- rMATS_filt3
rm(rMATS_filt3)
#### Abdomen DF ####
rmats_ab <- data.frame()
for (i in 1:length(rMATS_filt3ind_AB_DRY_vs_AB_WET)) {
temp <- rMATS_filt3ind_AB_DRY_vs_AB_WET[[i]] %>% select(ID, geneSymbol, IncFormLen, SkipFormLen, FDR, IncLevelDifference, event, comp)
rmats_ab <- rbind(rmats_ab, temp)
}
rmats_ab$event <- factor(rmats_ab$event, levels = events)
rmats_ab <- rmats_ab %>% rename(GeneID=geneSymbol)
rmats_ab_splice_summary <- rmats_ab %>% group_by(event) %>%
summarize(n_spl_events = n(),
genes = length(unique(GeneID)),
spl_event_per_gene = n_spl_events/genes,
n_sig_spl_events = sum(as.numeric(FDR) < 0.05 & abs(IncLevelDifference) > 0.05)) %>% ungroup %>%
mutate(total_spl_events = sum(n_spl_events) ,
total_sig_spl_events = sum(n_sig_spl_events),
expected_spl_events = (n_spl_events/total_spl_events)*total_sig_spl_events,
event = factor(event, levels = c("A3SS", "A5SS", "SE", "IR", "MXE")))
# load DS_ab_allgenes from nucleotideUpset.R
rmats_ab_summ <- rmats_ab %>%  group_by(GeneID,event) %>%
summarize(all_events = n(),
all_sig_events = sum(FDR < 0.05 & abs(IncLevelDifference) > 0.05, na.rm = T),
sig_DS_event = if_else(all_sig_events > 0, "DS", "AS")) %>%
full_join(data.frame(unique(na.omit(DS_ab_allgenes))))%>%
mutate(tissue = "Abdomen")
rmats_ab_summ$event <- factor(rmats_ab_summ$event, levels =c(levels(rmats_ab_summ$event), "None"))
no_events_ab <- rmats_ab_summ$GeneID[is.na(rmats_ab_summ$event)]
rmats_ab_summ$event[is.na(rmats_ab_summ$event)]<- "None"
rmats_ab_summ$all_events[rmats_ab_summ$event == "None"] <- 0
rmats_ab_summ$all_sig_events[rmats_ab_summ$event == "None"] <- 0
rmats_ab_summ$sig_DS_event[rmats_ab_summ$event == "None"] <- "None"
rmats_ab_summ2 <- rmats_ab_summ %>%  group_by(GeneID, tissue) %>%
summarize(total = sum(all_events),
total_sig = sum(all_sig_events),
sig_DS =if_else(total_sig > 0,"DS", if_else(total > 0, "AS", "None")))
#### Thorax DF ####
rmats_th <- data.frame()
for (i in 1:length(rMATS_filt3ind_TH_DRY_vs_TH_WET)) {
temp <- rMATS_filt3ind_TH_DRY_vs_TH_WET[[i]] %>% select(ID, geneSymbol, IncFormLen, SkipFormLen, FDR, IncLevelDifference, event, comp)
rmats_th <- rbind(rmats_th, temp)
}
rmats_th <- rmats_th %>% rename(GeneID=geneSymbol)
rmats_th$event <- factor(rmats_th$event, levels = events)
rmats_th_summ <- rmats_th %>% group_by(GeneID,event) %>%
summarize(all_events = n(),
all_sig_events = sum(FDR < 0.05 & abs(IncLevelDifference)>0.05, na.rm = T),
sig_DS_event = if_else(all_sig_events > 0, "DS", "AS")) %>%
full_join(data.frame(unique(na.omit(DS_th_allgenes))))%>%
mutate(tissue = "Thorax")
rmats_th_summ$event <- factor(rmats_th_summ$event, levels =c(levels(rmats_th_summ$event), "None"))
no_events_th <- rmats_th_summ$GeneID[is.na(rmats_th_summ$event)]
rmats_th_summ$event[is.na(rmats_th_summ$event)]<- "None"
rmats_th_summ$all_events[rmats_th_summ$event == "None"] <- 0
rmats_th_summ$all_sig_events[rmats_th_summ$event == "None"] <- 0
rmats_th_summ$sig_DS_event[rmats_th_summ$event == "None"] <- "None"
rmats_th_summ2 <- rmats_th_summ %>%  group_by(GeneID, tissue) %>%
summarize(total = sum(all_events),
total_sig = sum(all_sig_events),
sig_DS =if_else(total_sig > 0,"DS", if_else(total > 0, "AS", "None")))
rmats_summ <- rbind(rmats_ab_summ, rmats_th_summ) %>%
left_join(theta) %>%
left_join(omega) %>%
mutate(sig_DS_event = factor(sig_DS_event, levels = c("None", "AS", "DS")))
summary(rmats_summ)
fish.tests <- rmats_summ %>%
group_by(tissue, event) %>%
summarise(n_genes = length(unique(GeneID)),
n_all = sum(all_events),
n_sig = sum(all_sig_events)) %>%
group_by(tissue) %>%
mutate(sum_events = sum(n_all),
sum_sig_events = sum(n_sig)) %>%
group_by(tissue, event) %>%
mutate(exp_sig = round(sum_sig_events*(n_all/sum_events), 0),
fish.exact =  (rstatix::fisher_test(as.table(rbind(c(n_sig, n_all-n_sig), c(exp_sig, n_all-exp_sig)))))$p,
fish.adj = p.adjust(fish.exact, n = 5, method = "BH"))
fish.tests$fish.exact
knitr::kable(fish.tests)
rmats_summ2 <- rbind(rmats_ab_summ2, rmats_th_summ2) %>%
left_join(theta) %>%
left_join(omega) %>%
mutate(sig_DS = factor(sig_DS, levels = c("None", "AS", "DS")))
events <- c("A3SS","A5SS", "SE", "IR","MXE")
rmats_summ3 <- rmats_summ %>%
select(GeneID, tissue, event, sig_DS_event, pi_rel, omega) %>%
spread (key = "event", value = sig_DS_event) %>%
ungroup() %>%
rowwise(GeneID) %>%
mutate(is_DS = sum(c(A3SS == "DS", A5SS == "DS", SE == "DS", IR == "DS", MXE == "DS"), na.rm = T ),
is_AS = sum(c(A3SS == "AS", A5SS == "AS", SE == "AS", IR == "AS", MXE == "AS"), na.rm = T )) %>%
filter((is_DS == 1 | is_AS == 1 | None == "None")) %>%
filter((is_DS + is_AS <= 1)) %>%
select(c(1:10)) %>%
gather (-c(GeneID, tissue,  pi_rel, omega), key = event, value = sig_DS_event) %>%
mutate(event = factor(event, levels = c(events, "None"))) %>%
na.omit()
rmats_summ2 <- rmats_summ2 %>% mutate(sig_DS = factor(sig_DS, levels = c("None", "AS", "DS")))
rmats_summ3 <- rmats_summ3 %>% mutate(sig_DS_event = factor(sig_DS_event, levels = c("None", "AS", "DS")))
summary(rmats_summ2)
summary(rmats_summ3)
#### Summary Figures ####
n_events_plot <- rmats_summ2 %>% ggplot() +
geom_histogram(aes(x = total), bins = 100)+
labs(x = "Events", y = "Genes") +
theme_bw(base_size = 12) +
theme(panel.grid = element_blank()) +
facet_grid (~ tissue, scales = "free")
n_sigevents_plot <- rmats_summ2 %>% ggplot() +
geom_histogram(aes(x = total_sig), bins = 100)+
labs(x = "Significant events", y = "Genes") +
theme_bw(base_size = 12) +
theme(panel.grid = element_blank())+
facet_grid (~ tissue, scales = "free")
n_nonsigevents_plot <- rmats_summ2 %>% ggplot() +
geom_histogram(aes(x = total - total_sig), bins = 100) +
labs(x = "Nonsignificant events", y = "Genes") +
theme_bw(base_size = 12) +
theme(panel.grid = element_blank())+
facet_grid (~ tissue, scales = "free")
ggarrange(n_events_plot, n_sigevents_plot, n_nonsigevents_plot,
ncol = 1, nrow = 3, labels = "AUTO", align = "hv")
ggsave(filename = "/Users/rachelsteward/OneDrive/B_anynana_altSplicing/B_anynana_diff_expr_analysis/B_anynana_rMATS/B_anynana_rMATS_filt3_eventsingenes.pdf", height = 9, width = 6)
n_events_plot<- rmats_summ[rmats_summ$event != "None",] %>% ggplot() +
geom_histogram(aes(x = all_events), bins = 100)+
labs(x = "Events", y = "Genes") +
theme_bw(base_size = 12) +
theme(panel.grid = element_blank()) +
facet_grid (tissue ~ event, scales = "free")
n_sigevents_plot <- rmats_summ[rmats_summ$event != "None",] %>% ggplot() +
geom_histogram(aes(x = all_sig_events), bins = 15)+
labs(x = "Events", y = "Genes") +
theme_bw(base_size = 12) +
theme(panel.grid = element_blank()) +
facet_grid (tissue ~ event, scales = "free")
ggarrange(n_events_plot, n_sigevents_plot,
ncol = 1, nrow = 2, labels = "AUTO")
ggsave(filename = "/Users/rachelsteward/OneDrive/B_anynana_altSplicing/B_anynana_diff_expr_analysis/B_anynana_rMATS/B_anynana_rMATS_filt3_typesingenes.pdf", height =9, width = 9)
ls()
rm(DS_ab_allgenes, DS_th_allgenes, fish.tests, i, n_events_plot, n_nonsigevents_plot, n_sigevents_plot,
no_events_ab, no_events_th, rmats_ab_splice_summary, rmats_ab_summ, rmats_ab_summ2, rMATS_filt3ind_AB_DRY_vs_AB_WET,
rMATS_filt3ind_TH_DRY_vs_TH_WET, rmats_th_summ, rmats_th_summ2,  temp)
rmats_event_ab_pi_mod2 <- brm(pi_rel ~ event,
data=rmats_summ3[rmats_summ3$tissue == "Abdomen",],
family = gaussian(),
prior = c(prior(normal(0, 1), class = "Intercept"),
# set for all "b" coefficients
prior(normal(0, 1), class = "b")),
warmup = 500,
iter = 5000,
chains = 3)
save(rmats_event_ab_pi_mod2, file = "/Users/rachelsteward/OneDrive/B_anynana_altSplicing/PanicTime/rmats_event_ab_pi_mod2.Rdata" )rmats_event_th_pi_mod2 <- brm(pi_rel ~ event,
data=rmats_summ3[rmats_summ3$tissue == "Thorax",],
family = gaussian(),
prior = c(prior(normal(0, 1), class = "Intercept"),
# set for all "b" coefficients
prior(normal(0, 1), class = "b")),
warmup = 500,
iter = 5000,
chains = 3)
save(rmats_event_th_pi_mod2, file = "../rmats_event_th_pi_mod2.Rdata" )
rmats_event_ab_omega_mod2 <- brm(bf(sqrt(omega) ~ event , sigma ~ event),
data=rmats_summ3[rmats_summ3$omega < 5 & rmats_summ3$tissue == "Abdomen",],
family = skew_normal(),
prior = c(prior(normal(0, 1), class = "Intercept"),
# set for all "b" coefficients
prior(normal(0, 1), class = "b")),
warmup = 2000,
iter = 7000,
chains = 3)
save(rmats_event_ab_omega_mod2, file = "/Users/rachelsteward/OneDrive/B_anynana_altSplicing/PanicTime/rmats_event_ab_omega_mod2.Rdata" )
rmats_event_th_omega_mod2 <- brm(bf(sqrt(omega) ~ event , sigma ~ event),
data=rmats_summ3[rmats_summ3$omega < 5 & rmats_summ3$tissue == "Thorax",],
family = skew_normal(),
prior = c(prior(normal(0, 1), class = "Intercept"),
# set for all "b" coefficients
prior(normal(0, 1), class = "b")),
warmup = 2000,
iter = 7000,
chains = 3)
save(rmats_event_th_omega_mod2, file = "../rmats_event_th_omega_mod2.Rdata" )save(rmats_event_ab_pi_mod2, file = "/Users/rachelsteward/OneDrive/B_anynana_altSplicing/PanicTime/rmats_event_ab_pi_mod2.Rdata" )
save(rmats_event_ab_pi_mod2, file = "/Users/rachelsteward/OneDrive/B_anynana_altSplicing/PanicTime/rmats_event_ab_pi_mod2.Rdata" )
rmats_event_th_pi_mod2 <- brm(pi_rel ~ event,
data=rmats_summ3[rmats_summ3$tissue == "Thorax",],
family = gaussian(),
prior = c(prior(normal(0, 1), class = "Intercept"),
# set for all "b" coefficients
prior(normal(0, 1), class = "b")),
warmup = 500,
iter = 5000,
chains = 3)
save(rmats_event_th_pi_mod2, file = "../rmats_event_th_pi_mod2.Rdata" )
rbind(conditional_effects(rmats_event_ab_omega_mod2)[[1]][,5:9] %>%  mutate(response = "omega", tissue = "Abdomen"),
conditional_effects(rmats_event_ab_pi_mod2)[[1]][,4:8] %>%  mutate(response = "pi", tissue = "Abdomen"),
conditional_effects(rmats_event_th_omega_mod2)[[1]][,5:9] %>%  mutate(response = "omega", tissue = "Thorax"),
conditional_effects(rmats_event_th_pi_mod2)[[1]][,4:8] %>%  mutate(response = "pi", tissue = "Thorax")
) %>%
ggplot(aes(x = effect1__, y = estimate__)) +
geom_errorbar(aes(ymin = lower__, ymax = upper__), width = 0.25) +
geom_point() +
labs( x = NULL, y ="Estimate")+
theme_bw() +
facet_grid(response~ tissue, scales = "free")
save(rmats_event_th_pi_mod2, file = "../rmats_event_th_pi_mod2.Rdata" )
save(rmats_event_th_omega_mod2, file = "../rmats_event_th_omega_mod2.Rdata" )
save(rmats_event_ab_pi_mod2, file = "/Users/rachelsteward/OneDrive/B_anynana_altSplicing/PanicTime/rmats_event_ab_pi_mod2.Rdata" )
rmats_event_ab_omega_mod2 <- brm(bf(sqrt(omega) ~ event , sigma ~ event),
data=rmats_summ3[rmats_summ3$omega < 5 & rmats_summ3$tissue == "Abdomen",],
family = skew_normal(),
prior = c(prior(normal(0, 1), class = "Intercept"),
# set for all "b" coefficients
prior(normal(0, 1), class = "b")),
warmup = 2000,
iter = 7000,
chains = 3)
save(rmats_event_ab_omega_mod2, file = "/Users/rachelsteward/OneDrive/B_anynana_altSplicing/PanicTime/rmats_event_ab_omega_mod2.Rdata" )
conditional_effects(rmats_event_ab_omega_mod2)
rmats_event_DS_ab_pi_mod <- brm(pi_rel ~ event_DS_predict ,
data=rmats_summ[rmats_summ$tissue == "Abdomen",],
family = gaussian(),
prior = c(prior(normal(0, 1), class = "Intercept"),
# set for all "b" coefficients
prior(normal(0, 1), class = "b")),
warmup = 500,
iter = 5000,
chains = 3)
save(rmats_event_DS_ab_pi_mod, file = "/Users/rachelsteward/OneDrive/B_anynana_altSplicing/B_anynana_diff_expr_analysis/B_anynana_rMATS/rmats_event_DS_ab_pi_mod.Rdata")
rmats_summ$event_DS_predict <- factor(paste(rmats_summ$event, rmats_summ$sig_DS_event, sep = "_"))
rmats_event_DS_ab_pi_mod <- brm(pi_rel ~ event_DS_predict ,
data=rmats_summ[rmats_summ$tissue == "Abdomen",],
family = gaussian(),
prior = c(prior(normal(0, 1), class = "Intercept"),
# set for all "b" coefficients
prior(normal(0, 1), class = "b")),
warmup = 500,
iter = 5000,
chains = 3)
save(rmats_event_DS_ab_pi_mod, file = "/Users/rachelsteward/OneDrive/B_anynana_altSplicing/B_anynana_diff_expr_analysis/B_anynana_rMATS/rmats_event_DS_ab_pi_mod.Rdata")
# rmats_event_DS_th_pi_mod <- brm(pi_rel ~ event_DS_predict ,
#                              data=rmats_summ[rmats_summ$tissue == "Thorax",],
#                              family = gaussian(),
#                              prior = c(prior(normal(0, 1), class = "Intercept"),
#                                        # set for all "b" coefficients
#                                        prior(normal(0, 1), class = "b")),
#                              warmup = 500,
#                              iter = 5000,
#                              chains = 3)
# save(rmats_event_DS_th_pi_mod, file = "/Users/rachelsteward/OneDrive/B_anynana_altSplicing/B_anynana_diff_expr_analysis/B_anynana_rMATS/rmats_event_DS_th_pi_mod.Rdata")
rmats_event_DS_th_pi_mod <- brm(pi_rel ~ event_DS_predict ,
data=rmats_summ[rmats_summ$tissue == "Thorax",],
family = gaussian(),
prior = c(prior(normal(0, 1), class = "Intercept"),
# set for all "b" coefficients
prior(normal(0, 1), class = "b")),
warmup = 500,
iter = 5000,
chains = 3)
save(rmats_event_DS_th_pi_mod, file = "/Users/rachelsteward/OneDrive/B_anynana_altSplicing/B_anynana_diff_expr_analysis/B_anynana_rMATS/rmats_event_DS_th_pi_mod.Rdata")
rbind(conditional_effects(rmats_ab_omega_mod2)[[1]][,5:9] %>%  mutate(response = "omega", tissue = "Abdomen"),
conditional_effects(rmats_ab_pi_mod2)[[1]][,4:8] %>%  mutate(response = "pi", tissue = "Abdomen"),
conditional_effects(rmats_th_omega_mod2)[[1]][,5:9] %>%  mutate(response = "omega", tissue = "Thorax"),
conditional_effects(rmats_th_pi_mod2)[[1]][,4:8] %>%  mutate(response = "pi", tissue = "Thorax")
) %>%
ggplot(aes(x = effect1__, y = estimate__)) +
geom_errorbar(aes(ymin = lower__, ymax = upper__), width = 0.25) +
geom_point() +
labs( x = NULL, y ="Estimate")+
theme_bw() +
facet_grid(response~ tissue, scales = "free")
ggarrange(ggarrange(NULL, NULL),
ggarrange( rmats_events_pi_ab, rmats_events_pi_th, ncol = 2, align ="hv", labels = c("A    Abdomen", "B    Thorax"), vjust = 0, hjust = 0),
ggarrange(rmats_event_ab_pi_mod2_hyp, rmats_event_th_pi_mod2_hyp, ncol = 2, align ="hv"),
ggarrange(NULL, NULL),
ggarrange(  rmats_events_omega_ab, rmats_events_omega_th, ncol = 2, align ="hv",  labels = c("C", "D"), vjust = 0, hjust = 0),
ggarrange(rmats_event_ab_omega_mod2_hyp, rmats_event_th_omega_mod2_hyp,)
)
rbind(conditional_effects(rmats_event_ab_omega_mod2)[[1]][,5:9] %>%  mutate(response = "omega", tissue = "Abdomen"),
conditional_effects(rmats_event_ab_pi_mod2)[[1]][,4:8] %>%  mutate(response = "pi", tissue = "Abdomen"),
conditional_effects(rmats_event_th_omega_mod2)[[1]][,5:9] %>%  mutate(response = "omega", tissue = "Thorax"),
conditional_effects(rmats_event_th_pi_mod2)[[1]][,4:8] %>%  mutate(response = "pi", tissue = "Thorax")
) %>%
ggplot(aes(x = effect1__, y = estimate__)) +
geom_errorbar(aes(ymin = lower__, ymax = upper__), width = 0.25) +
geom_point() +
labs( x = NULL, y ="Estimate")+
theme_bw() +
facet_grid(response~ tissue, scales = "free")
conditional_effects(rmats_event_ab_omega_mod2)
knitr::opts_chunk$set(echo = TRUE, warnings = FALSE, message = FALSE)
list.of.packages <- c("tidyverse",
"ggpubr",
"ggsci",
"ComplexHeatmap",
"circlize",
"simplifyEnrichment",
"grid")
new.packages <- list.of.packages[!(list.of.packages %in% installed.packages()[,"Package"])]
new.packages
if(length(new.packages)) install.packages(new.packages)
library(tidyverse)
library(ggpubr)
library(ComplexHeatmap)
library(ggsci)
library(circlize)
library(simplifyEnrichment)
library(grid)
mypal_gsea<- c("#7876B1","#0072B5","#BC3C29", "white")
myPath <- "/Users/rachelsteward/OneDrive/B_anynana_altSplicing/PanicTime/genelists/"
GOterm_Upset_list <- list()
files = list.files(path = paste0(myPath, "REVIGO_S/"), pattern = "BP.GSEA_result_parentchild.REVIGO.tsv")
for ( i in 1:length(files)){
temp = read_tsv(paste0(myPath, "REVIGO_S/", files[i]), col_names = c("GOterm", "P_value")) %>%
mutate(log10_P = 0-log10(P_value), geneset = gsub(".h", "", str_sub(files[i], 1, 9))) %>%
arrange(log10_P) %>% filter(P_value <0.05)
GOterm_Upset_list[[i]] <- temp$GOterm
names(GOterm_Upset_list)[i] <- str_split(files[i], pattern = "_", n = 3)[[1]][2]
}
files = list.files(path = paste0(myPath, "REVIGO_F/"), pattern = "BP.GSEA_result_parentchild.REVIGO.tsv")
for ( i in 1:length(files)){
temp = read_tsv(paste0(myPath, "REVIGO_F/", files[i]), col_names = c("GOterm", "P_value")) %>%
mutate(log10_P = 0-log10(P_value), geneset = gsub(".h", "", str_sub(files[i], 1, 9))) %>%
arrange(log10_P) %>% filter(P_value <0.05)
GOterm_Upset_list[[i+6]] <- temp$GOterm
names(GOterm_Upset_list)[i+6] <- str_split(files[i], pattern = "_", n = 3)[[1]][2]
}
files = list.files(path = paste0(myPath, "REVIGO_I/"), pattern = "BP.GSEA_result_parentchild.REVIGO.tsv")
for ( i in 1:length(files)){
temp = read_tsv(paste0(myPath, "REVIGO_I/", files[i]), col_names = c("GOterm", "P_value")) %>%
mutate(log10_P = 0-log10(P_value), geneset = gsub(".h", "", str_sub(files[i], 1, 9))) %>%
arrange(log10_P) %>% filter(P_value <0.05)
GOterm_Upset_list[[i+12]] <- temp$GOterm
names(GOterm_Upset_list)[i+12] <- str_split(files[i], pattern = "_", n = 3)[[1]][2]
}
summary(GOterm_Upset_list)
gseaUpSet <- function(m, ...){
cs = comb_size(m)
topAnno = HeatmapAnnotation(
"Intersections" = anno_barplot(cs, gp = gpar(fill = c("gray", "gray", "gray", mypal_gsea[c(2,1,3)])),
height = unit(3, "cm"), ylim = c(0,max(cs)+100 ),
axis_param = list(gp = gpar(fontsize = 10)),
border = FALSE),
annotation_name_side = "left",
annotation_label = gt_render(c("Intersections")),
gap = unit(2, "mm"))
rightAnno =  upset_right_annotation(
m,
axis_param = list( labels_rot = 0),
width = unit(2, "cm"),
annotation_label = gt_render(c("Set")),
gp = gpar(fill = mypal_gsea[c(2,1,3)]))
ups = UpSet(m, set_order = c( "DS","DEDS", "DE"), top_annotation = topAnno, right_annotation = rightAnno)
ht = draw(ups)
od = column_order(ht)
decorate_annotation("Intersections", {
grid.text(cs[od], x = seq_along(cs), y = unit(cs[od], "native") + unit(2, "pt"),
default.units = "native", just = c("bottom"),
gp = gpar(fontsize = 10, col = "#404040"), rot = 0)
})
return(invisible(NULL))
}
GOterm_comb_mat = normalize_comb_mat(list(ab_season_matrix = make_comb_mat(GOterm_Upset_list[c(1:3)]), # abdomen ds effects plot
th_season_matrix = make_comb_mat(GOterm_Upset_list[c(4:6)]),
ab_family_matrix = make_comb_mat(GOterm_Upset_list[c(7:9)]),
th_family_matrix = make_comb_mat(GOterm_Upset_list[c(10:12)]),
ab_sxf_matrix = make_comb_mat(GOterm_Upset_list[c(13:14)]),
th_sxf_matrix = make_comb_mat(GOterm_Upset_list[c(15:16)])))
knitr::kable(sapply(GOterm_comb_mat, comb_size), caption = "Combination matrix of GO terms (position 1 = DE, 2 = DEDS, 3=DS)")
gsea_BPUpset_plots <- list()
gsea_BPUpset_plots[[1]] <- gseaUpSet(GOterm_comb_mat$ab_season_matrix)
gsea_BPUpset_plots[[2]] <- gseaUpSet(GOterm_comb_mat$th_season_matrix)
gsea_BPUpset_plots[[3]] <- gseaUpSet(GOterm_comb_mat$ab_family_matrix)
gsea_BPUpset_plots[[4]] <- gseaUpSet(GOterm_comb_mat$th_family_matrix)
gsea_BPEuler_plots <- list()
gsea_BPEuler_plots[[1]] <- plot(eulerr::euler(GOterm_Upset_list[c(1:3)]), quantities = TRUE, fills = adjustcolor(mypal_gsea[c(2,1,3)], alpha = 0.75), labels = c("DE only", "DE & DS", "DS only"))
gsea_BPEuler_plots[[2]] <-   plot(eulerr::euler(GOterm_Upset_list[c(4:6)]), quantities = TRUE, fills = adjustcolor(mypal_gsea[c(2,1,3)], alpha = 0.75), labels =  c("DE only", "DE & DS", "DS only"))
gsea_BPEuler_plots[[3]] <-   plot(eulerr::euler(GOterm_Upset_list[c(7:9)]), quantities = TRUE, fills = adjustcolor(mypal_gsea[c(2,1,3)], alpha = 0.75), labels = c("DE only", "DE & DS", "DS only"))
gsea_BPEuler_plots[[4]] <-   plot(eulerr::euler(GOterm_Upset_list[c(10:12)]), quantities = TRUE, fills = adjustcolor(mypal_gsea[c(2,1,3)], alpha = 0.75), labels =  c("DE only", "DE & DS", "DS only"))
gsea_BPEuler_plots[[5]] <-   plot(eulerr::euler(GOterm_Upset_list[c(13:14)]), quantities = TRUE, fills = adjustcolor(mypal_gsea[c(2,3)], alpha = 0.75), labels =  c("DE only",  "DS only"))
gsea_BPEuler_plots[[6]] <-   plot(eulerr::euler(GOterm_Upset_list[c(15:16)]), quantities = TRUE, fills = adjustcolor(mypal_gsea[c(2,3)], alpha = 0.75), labels =  c("DE only",  "DS only"))
save(gsea_BPEuler_plots,file = "Data_in/gsea_BPEuler_plots.Rdata")
files = list.files(path = paste0(myPath, "REVIGO_S/SemSpace/"), pattern = "S_BP.REVIGO.csv")
topGO_BP <- data.frame(geneset = c( gsub("_B", "", str_sub(files[i], 1, 9))))
top_genes_S <- list()
for ( i in 1:length(files)){
temp = read_csv(paste0(myPath, "REVIGO_S/SemSpace/", files[i])) %>%
mutate(log10_P = 0-`log10 p-value`, geneset = gsub("_B", "", str_sub(files[i], 1, 9))) %>%
arrange(log10_P) %>%
dplyr::select(geneset, term_ID, description, log10_P, dispensability, eliminated)
topGO_BP <- full_join(topGO_BP, temp)
top_genes_S[i]<- temp %>%filter(log10_P > -log10(0.05)) %>% dplyr::select(term_ID)
names(top_genes_S)[i] <- gsub("_B", "", str_sub(files[i], 1, 9))
}
top_genes_F <- list()
files = list.files(path = paste0(myPath, "REVIGO_F/SemSpace/"), pattern = "F_BP.REVIGO.csv")
for ( i in 1:length(files)){
temp = read_csv(paste0(myPath, "REVIGO_F/SemSpace/", files[i])) %>%
mutate(log10_P = 0-Value, geneset = gsub("_B", "", str_sub(files[i], 1, 9))) %>%
arrange(log10_P) %>%
dplyr::select(geneset, term_ID=TermID, description=Name, log10_P, dispensability=Dispensability, eliminated=Eliminated)
topGO_BP <- full_join(topGO_BP, temp)
top_genes_F[i]<- temp %>%filter(log10_P > -log10(0.05)) %>% dplyr::select(term_ID)
names(top_genes_F)[i] <- gsub("_B", "", str_sub(files[i], 1, 9))
}
top_genes_I <- list()
files = list.files(path = paste0(myPath, "REVIGO_I/SemSpace/"), pattern = "I_BP.REVIGO.csv")
for ( i in 1:length(files)){
temp = read_csv(paste0(myPath, "REVIGO_I/SemSpace/", files[i])) %>%
mutate(log10_P = 0-`log10 p-value`, geneset = gsub("_B", "", str_sub(files[i], 1, 9))) %>%
arrange(log10_P) %>%
dplyr::select(geneset, term_ID, description, log10_P, dispensability, eliminated)
topGO_BP <- full_join(topGO_BP, temp)
top_genes_I[i]<- temp %>%filter(log10_P > -log10(0.05)) %>% dplyr::select(term_ID)
names(top_genes_I)[i] <- gsub("_B", "", str_sub(files[i], 1, 9))
}
term_IDs<- unique(topGO_BP$term_ID) %>% na.omit()
functions_df <- data.frame(term_ID = c(), description = c())
for (i in 1:length(term_IDs)){
temp <- topGO_BP %>% filter(term_ID == term_IDs[[i]]) %>% dplyr::select(term_ID, description) %>% head(1)
functions_df <- rbind(functions_df, temp)
}
write_csv(functions_df,"B_anynana_BP_enrichedFunctions.csv")
topGO_BP_wide <- topGO_BP %>%  na.omit() %>%
dplyr::select(geneset,term_ID, log10_P) %>%
left_join(functions_df) %>%
mutate(GOterm = paste0(term_ID,"-", description)) %>%
dplyr::select(term_ID, GOterm,geneset, log10_P) %>%
pivot_wider(names_from = geneset, values_from = log10_P)
summary(topGO_BP)
summary(top_genes_S)
summary(top_genes_F)
summary(top_genes_I)
summary(GOterm_Upset_list)
summary(topGO_BP_wide)
GOterm_Upset_list <- list()
files = list.files(path = paste0(myPath, "REVIGO_S/"), pattern = "BP.GSEA_result_parentchild.REVIGO.tsv")
for ( i in 1:length(files)){
temp = read_tsv(paste0(myPath, "REVIGO_S/", files[i]), col_names = c("GOterm", "P_value")) %>%
mutate(log10_P = 0-log10(P_value), geneset = gsub(".h", "", str_sub(files[i], 1, 9))) %>%
arrange(log10_P) %>% filter(P_value <0.05)
GOterm_Upset_list[[i]] <- temp$GOterm
names(GOterm_Upset_list)[i] <- str_split(files[i], pattern = "_", n = 3)[[1]][2]
}
temp = read_tsv(paste0(myPath, "REVIGO_S/", files[i]), col_names = c("GOterm", "P_value"))
head(temp)
