diff --git a/.Rhistory b/.Rhistory
index e0ec83a..f97a208 100644
--- a/.Rhistory
+++ b/.Rhistory
@@ -1,363 +1,3 @@
-ysidehistogram.args = list(fill="blue", color="white"),
-point.args = list(alpha=.1)) +
-ggplot2::geom_density_2d_filled(alpha=.7) +
-ggplot2::scale_fill_manual(values = c(ggplot2::alpha("white",0),
-ggplot2::alpha(pals::gnuplot(8)[-1],.5)
-))
-# plots[["all"]] <-
-res2 |>
-ggstatsplot::ggscatterstats(x="p_HumanCellLandscape",
-y="p_DescartesHuman",
-xsidehistogram.args = list(fill="red", color="white"),
-ysidehistogram.args = list(fill="blue", color="white"),
-point.args = list(alpha=.1)) +
-ggplot2::geom_density_2d_filled(alpha=.7) +
-ggplot2::scale_fill_manual(values = c(ggplot2::alpha("white",0),
-ggplot2::alpha(pals::gnuplot(10)[-1],.5)
-))
-make_density_plot <- function(res,
-x="p_HumanCellLandscape",
-y="p_DescartesHuman"){
-res |>
-ggstatsplot::ggscatterstats(x=X,
-y=y,
-xsidehistogram.args = list(fill="red", color="white"),
-ysidehistogram.args = list(fill="blue", color="white"),
-point.args = list(alpha=.1)) +
-ggplot2::geom_density_2d_filled(alpha=.7) +
-ggplot2::scale_fill_manual(values = c(ggplot2::alpha("white",0),
-ggplot2::alpha(pals::gnuplot(8)[-1],.5)
-))
-}
-plots[["all"]] <- make_density_plot(res2)
-make_density_plot <- function(res,
-x="p_HumanCellLandscape",
-y="p_DescartesHuman"){
-res |>
-ggstatsplot::ggscatterstats(x=x,
-y=y,
-xsidehistogram.args = list(fill="red", color="white"),
-ysidehistogram.args = list(fill="blue", color="white"),
-point.args = list(alpha=.1)) +
-ggplot2::geom_density_2d_filled(alpha=.7) +
-ggplot2::scale_fill_manual(values = c(ggplot2::alpha("white",0),
-ggplot2::alpha(pals::gnuplot(8)[-1],.5)
-))
-}
-plots[["all"]] <- make_density_plot(res2)
-make_density_plot <- function(res,
-x="p_HumanCellLandscape",
-y="p_DescartesHuman"){
-res |>
-ggstatsplot::ggscatterstats(x=x,
-y=y,
-xsidehistogram.args = list(fill="red", color="white"),
-ysidehistogram.args = list(fill="blue", color="white"),
-point.args = list(alpha=.1)) +
-ggplot2::geom_density_2d_filled(alpha=.7) +
-ggplot2::scale_fill_manual(values = c(ggplot2::alpha("white",0),
-ggplot2::alpha(pals::gnuplot(8)[-1],.5)
-))
-}
-plots[["all"]] <- make_density_plot(res2)
-res |>
-ggstatsplot::ggscatterstats(x=get(),
-y=y,
-xsidehistogram.args = list(fill="red", color="white"),
-ysidehistogram.args = list(fill="blue", color="white"),
-point.args = list(alpha=.1)) +
-ggplot2::geom_density_2d_filled(alpha=.7) +
-ggplot2::scale_fill_manual(values = c(ggplot2::alpha("white",0),
-ggplot2::alpha(pals::gnuplot(8)[-1],.5)
-))
-make_density_plot <- function(res,
-x="p_HumanCellLandscape",
-y="p_DescartesHuman"){
-res |>
-ggstatsplot::ggscatterstats(x=get(),
-y=y,
-xsidehistogram.args = list(fill="red", color="white"),
-ysidehistogram.args = list(fill="blue", color="white"),
-point.args = list(alpha=.1)) +
-ggplot2::geom_density_2d_filled(alpha=.7) +
-ggplot2::scale_fill_manual(values = c(ggplot2::alpha("white",0),
-ggplot2::alpha(pals::gnuplot(8)[-1],.5)
-))
-}
-make_density_plot <- function(res,
-x="p_HumanCellLandscape",
-y="p_DescartesHuman"){
-res |>
-ggstatsplot::ggscatterstats(x=get(),
-y=y,
-xsidehistogram.args = list(fill="red", color="white"),
-ysidehistogram.args = list(fill="blue", color="white"),
-point.args = list(alpha=.1)) +
-ggplot2::geom_density_2d_filled(alpha=.7) +
-ggplot2::scale_fill_manual(values = c(ggplot2::alpha("white",0),
-ggplot2::alpha(pals::gnuplot(8)[-1],.5)
-))
-}
-plots[["all"]] <- make_density_plot(res2)
-make_density_plot <- function(res,
-x="p_HumanCellLandscape",
-y="p_DescartesHuman"){
-res |>
-ggstatsplot::ggscatterstats(x=get(x),
-y=y,
-xsidehistogram.args = list(fill="red", color="white"),
-ysidehistogram.args = list(fill="blue", color="white"),
-point.args = list(alpha=.1)) +
-ggplot2::geom_density_2d_filled(alpha=.7) +
-ggplot2::scale_fill_manual(values = c(ggplot2::alpha("white",0),
-ggplot2::alpha(pals::gnuplot(8)[-1],.5)
-))
-}
-plots[["all"]] <- make_density_plot(res2)
-res |>
-ggstatsplot::ggscatterstats(x=!!ggplot2::sym(x),
-y=y,
-xsidehistogram.args = list(fill="red", color="white"),
-ysidehistogram.args = list(fill="blue", color="white"),
-point.args = list(alpha=.1)) +
-ggplot2::geom_density_2d_filled(alpha=.7) +
-ggplot2::scale_fill_manual(values = c(ggplot2::alpha("white",0),
-ggplot2::alpha(pals::gnuplot(8)[-1],.5)
-))
-make_density_plot <- function(res,
-x="p_HumanCellLandscape",
-y="p_DescartesHuman"){
-res |>
-ggstatsplot::ggscatterstats(x=!!ggplot2::sym(x),
-y=y,
-xsidehistogram.args = list(fill="red", color="white"),
-ysidehistogram.args = list(fill="blue", color="white"),
-point.args = list(alpha=.1)) +
-ggplot2::geom_density_2d_filled(alpha=.7) +
-ggplot2::scale_fill_manual(values = c(ggplot2::alpha("white",0),
-ggplot2::alpha(pals::gnuplot(8)[-1],.5)
-))
-}
-plots[["all"]] <- make_density_plot(res2)
-make_density_plot <- function(res,
-x="p_HumanCellLandscape",
-y="p_DescartesHuman"){
-res |>
-ggstatsplot::ggscatterstats(x=!!ggplot2::sym(x),
-y=!!ggplot2::sym(y),
-xsidehistogram.args = list(fill="red", color="white"),
-ysidehistogram.args = list(fill="blue", color="white"),
-point.args = list(alpha=.1)) +
-ggplot2::geom_density_2d_filled(alpha=.7) +
-ggplot2::scale_fill_manual(values = c(ggplot2::alpha("white",0),
-ggplot2::alpha(pals::gnuplot(8)[-1],.5)
-))
-}
-plots[["all"]] <- make_density_plot(res2)
-plots[["all"]]
-make_density_plot <- function(res,
-x="p_HumanCellLandscape",
-y="p_DescartesHuman"){
-res |>
-ggstatsplot::ggscatterstats(x=!!ggplot2::sym(x),
-y=!!ggplot2::sym(y),
-xsidehistogram.args = list(fill="magenta", color="white"),
-ysidehistogram.args = list(fill="blue", color="white"),
-point.args = list(alpha=.1)) +
-ggplot2::geom_density_2d_filled(alpha=.7) +
-ggplot2::scale_fill_manual(values = c(ggplot2::alpha("white",0),
-ggplot2::alpha(pals::gnuplot(8)[-1],.5)
-))
-}
-plots[["all"]] <- make_density_plot(res2)
-plots[["significant"]] <- make_density_plot(res_sig)
-plots[["significant"]]
-make_density_plot <- function(res,
-x="p_HumanCellLandscape",
-y="p_DescartesHuman"){
-res |>
-ggstatsplot::ggscatterstats(x=!!ggplot2::sym(x),
-y=!!ggplot2::sym(y),
-xsidehistogram.args = list(fill="magenta", color="white"),
-ysidehistogram.args = list(fill="blue", color="white"),
-point.args = list(alpha=.1)) +
-ggplot2::geom_density_2d_filled(alpha=.7) +
-ggplot2::scale_fill_manual(values = c(ggplot2::alpha("white",0),
-ggplot2::alpha(pals::gnuplot(15)[-1],.5)
-))
-}
-plots[["all"]] <- make_density_plot(res2)
-plots[["significant"]] <- make_density_plot(res_sig)
-plots[["all"]]
-plots[["significant"]]
-plots[["significant"]]  +
-ggplot2::scale_x_log10() +
-ggplot2::scale_y_log10()
-make_density_plot <- function(res,
-x="p_HumanCellLandscape",
-y="p_DescartesHuman",
-log_vars=FALSE){
-p<- res |>
-ggstatsplot::ggscatterstats(x=!!ggplot2::sym(x),
-y=!!ggplot2::sym(y),
-xsidehistogram.args = list(fill="magenta", color="white"),
-ysidehistogram.args = list(fill="blue", color="white"),
-point.args = list(alpha=.1)) +
-ggplot2::geom_density_2d_filled(alpha=.7) +
-ggplot2::scale_fill_manual(values = c(ggplot2::alpha("white",0),
-ggplot2::alpha(pals::gnuplot(15)[-1],.5)
-))
-if(log_vars){
-p <- p+
-ggplot2::scale_x_log10() +
-ggplot2::scale_y_log10()
-}
-return(p)
-}
-add_logfc(results)
-res2 <- results|>
-# subset(stage=="Fetus") |>
-# subset(q<0.05)|>
-data.table::dcast.data.table(
-formula = hpo_id+cl_name ~ ctd,
-fun.aggregate = mean,
-drop = TRUE,
-value.var = c("p","q","logFC","estimate"))
-value.var <- intersect(c("p","q","logFC","estimate"),
-names(results))
-res2 <- results|>
-# subset(stage=="Fetus") |>
-# subset(q<0.05)|>
-data.table::dcast.data.table(
-formula = hpo_id+cl_name ~ ctd,
-fun.aggregate = mean,
-drop = TRUE,
-value.var = value.var)
-res2 <- res2[complete.cases(res2)][,test_id:=.I]
-### All results
-message(length(unique(res2$cl_name))," comparable celltypes.")
-message(length(unique(res2$hpo_id))," comparable phenotypes.")
-### Significant results in both CTDs
-res_sig <- res2[q_HumanCellLandscape<.05 & q_DescartesHuman<.05]
-message(length(unique(res_sig$cl_name))," comparable celltypes (FDR<0.05).")
-message(length(unique(res_sig$hpo_id))," comparable phenotypes (FDR<0.05).")
-plots <- list()
-make_density_plot <- function(res,
-x="p_HumanCellLandscape",
-y="p_DescartesHuman",
-log_vars=FALSE){
-p<- res |>
-ggstatsplot::ggscatterstats(x=!!ggplot2::sym(x),
-y=!!ggplot2::sym(y),
-xsidehistogram.args = list(fill="magenta", color="white"),
-ysidehistogram.args = list(fill="blue", color="white"),
-point.args = list(alpha=.1)) +
-ggplot2::geom_density_2d_filled(alpha=.7) +
-ggplot2::scale_fill_manual(values = c(ggplot2::alpha("white",0),
-ggplot2::alpha(pals::gnuplot(15)[-1],.5)
-))
-if(log_vars){
-p <- p+
-ggplot2::scale_x_log10() +
-ggplot2::scale_y_log10()
-}
-return(p)
-}
-plots[["all"]] <- make_density_plot(res2)
-plots[["significant"]] <- make_density_plot(res_sig,
-x="logFC_HumanCellLandscape",
-y="logFC_DescartesHuman",
-log_vars=TRUE)
-plots[["significant"]] <- make_density_plot(res_sig,
-x="logFC_HumanCellLandscape",
-y="logFC_DescartesHuman",
-log_vars=F)
-plots[["significant"]]
-plots[["logFC.all"]] <- make_density_plot(res2,
-x="logFC_HumanCellLandscape",
-y="logFC_DescartesHuman")
-plots[["logFC.all"]]
-plots[["logFC.all"]] <- make_density_plot(res2[!is.na(logFC)],
-x="logFC_HumanCellLandscape",
-y="logFC_DescartesHuman")
-res2 <- res2[complete.cases(res2)][,test_id:=.I]
-make_density_plot <- function(res,
-x="p_HumanCellLandscape",
-y="p_DescartesHuman",
-log_vars=FALSE){
-res <- res[!is.na(get(x)) & !is.na(get(y))]
-p<- res |>
-ggstatsplot::ggscatterstats(x=!!ggplot2::sym(x),
-y=!!ggplot2::sym(y),
-xsidehistogram.args = list(fill="magenta", color="white"),
-ysidehistogram.args = list(fill="blue", color="white"),
-point.args = list(alpha=.1)) +
-ggplot2::geom_density_2d_filled(alpha=.7) +
-ggplot2::scale_fill_manual(values = c(ggplot2::alpha("white",0),
-ggplot2::alpha(pals::gnuplot(15)[-1],.5)
-))
-if(log_vars){
-p <- p+
-ggplot2::scale_x_log10() +
-ggplot2::scale_y_log10()
-}
-return(p)
-}
-plots[["logFC.all"]] <- make_density_plot(res2,
-x="logFC_HumanCellLandscape",
-y="logFC_DescartesHuman")
-plots[["logFC.all"]]
-plots[["logFC.significant"]] <- make_density_plot(res_sig,
-x="logFC_HumanCellLandscape",
-y="logFC_DescartesHuman")
-plots[["logFC.significant"]]
-unique(results$ctd)[2]
-ctd_select=unique(results$ctd)[2]
-results=load_example_results()[ctd==ctd_select]
-results
-results <- map_celltype(results)
-add_logfc(results)
-value.var <- intersect(c("p","q","logFC","estimate"),
-names(results))
-comparison_var="stage"
-unique(results$stage)
-filters=list("stage"=c("Fetus","Adult"))
-results <- KGExplorer::filter_dt(results, filters = filters)
-comparison_var
-res2 <- results |>
-data.table::dcast.data.table(
-formula = as.formula(paste0("hpo_id+cl_name~",comparison_var)),
-fun.aggregate = mean,
-drop = TRUE,
-value.var = value.var)
-res2 <- res2[complete.cases(res2)][,test_id:=.I]
-### All results
-message(length(unique(res2$cl_name))," comparable celltypes.")
-message(length(unique(res2$hpo_id))," comparable phenotypes.")
-### Significant results in both CTDs
-res_sig <- RES_HCL[q_Adult<.05 & q_Fetus<.05]
-message(length(unique(res_sig$cell_ontology_mapped))," comparable celltypes (FDR<0.05).")
-message(length(unique(res_sig$cl_name))," comparable celltypes (FDR<0.05).")
-message(length(unique(res_sig$hpo_id))," comparable phenotypes (FDR<0.05).")
-plots <- list()
-filters[[comparison_var]]
-filters[[comparison_var]][1]
-x_var=filters[[comparison_var]][1]
-y_var=filters[[comparison_var]][2]
-data_stats <- lapply(plots, get_ggstatsplot_stats)
-source("~/Desktop/Rare Disease Celltyping/MSTExplorer/R/get_ggstatsplot_stats.R")
-data_stats <- lapply(plots, get_ggstatsplot_stats)
-data_stats
-plots
-source("~/Desktop/Rare Disease Celltyping/MSTExplorer/R/plot_density_cor.R")
-### Significant results in both CTDs
-res_sig <- res2[q_Adult<.05 & q_Fetus<.05]
-celltype_var="cl_name"
-comparison_var="stage"
-celltype_var="cl_name"
-group_values <- unique(results[[comparison_var]])
-group_values
-### Significant results in both CTDs
 res_sig <- res2[get(paste0("q_",group_values[2]))<qvalue_threshold &
 get(paste0("q_",group_values[2]))<qvalue_threshold]
 qvalue_threshold=0.05
@@ -510,3 +150,363 @@ devtools::check_man()
 devtools::check_man()
 devtools::check_man()
 library(MSTExplorer)
+devtools::check_man()
+library(MSTExplorer)
+devtools::check_man()
+library(MSTExplorer)
+prioritise_targets_out <- MSTExplorer::prioritise_targets(
+keep_deaths=NULL,
+keep_onsets=NULL,
+keep_specificity_quantiles = seq(30,40), ## NULL:70, 30-40:64
+keep_mean_exp_quantiles = seq(30,40), ## NULL:65, 10:55
+info_content_threshold=8, ## 8:55, 5:64
+effect_threshold=NULL, ## 1:39
+severity_score_gpt_threshold=NULL, ## 10:78, NULL:82
+symptom_intersection_threshold=.25, ## .25:57
+evidence_score_threshold=3,  ## 5:47, 4:47, 3:64
+top_n = 10, ## 5:38, 20:42, 30:45, 40:52, 50:55
+group_vars = "hpo_id")
+ttdi <- KGExplorer::get_ttd()
+#### Remove results that can't be linked to specific genes #####
+dat_sub <- ttdi$merged[!is.na(TARGETID) &
+!is.na(GENENAME2) &
+GENENAME2!="",]
+#### Filter by drug type ####
+if(!is.null(drug_types)){
+dat_sub <- dat_sub[
+grepl(paste(drug_types,collapse = "|"),DRUGNAME,ignore.case = TRUE) |
+grepl(paste(drug_types,collapse = "|"),DRUGTYPE,ignore.case = TRUE),]
+}
+devoptera::args2vars()
+#### Remove results that can't be linked to specific genes #####
+dat_sub <- ttdi$merged[!is.na(TARGETID) &
+!is.na(GENENAME2) &
+GENENAME2!="",]
+#### Filter by drug type ####
+if(!is.null(drug_types)){
+dat_sub <- dat_sub[
+grepl(paste(drug_types,collapse = "|"),DRUGNAME,ignore.case = TRUE) |
+grepl(paste(drug_types,collapse = "|"),DRUGTYPE,ignore.case = TRUE),]
+}
+#### Filter by status ####
+if(!is.null(keep_status)){
+dat_sub <- dat_sub[HIGHEST_STATUS %in% keep_status,]
+}
+if(!is.null(remove_status)){
+dat_sub <- dat_sub[!HIGHEST_STATUS %in% remove_status,]
+}
+dat_sub[,failed:=HIGHEST_STATUS %in% failed_status]
+#### Filter to only those in top_targets ####
+dat_sub2 <- (merge(
+dat_sub[failed==FALSE],
+top_targets,
+allow.cartesian = allow.cartesian,
+by.x = "GENENAME3",
+by.y = "gene_symbol")[,c("GENENAME2","TARGETID","TARGNAME",
+"INDICATI","DRUGID","DRUGNAME",
+"HIGHEST_STATUS",
+"disease_name","disease_id","hpo_name",
+"CellType","ontLvl")] |>
+unique())
+top_targets <- prioritise_targets_out$top_targets
+#### Filter by status ####
+if(!is.null(keep_status)){
+dat_sub <- dat_sub[HIGHEST_STATUS %in% keep_status,]
+}
+if(!is.null(remove_status)){
+dat_sub <- dat_sub[!HIGHEST_STATUS %in% remove_status,]
+}
+dat_sub[,failed:=HIGHEST_STATUS %in% failed_status]
+#### Filter to only those in top_targets ####
+dat_sub2 <- (merge(
+dat_sub[failed==FALSE],
+top_targets,
+allow.cartesian = allow.cartesian,
+by.x = "GENENAME3",
+by.y = "gene_symbol")[,c("GENENAME2","TARGETID","TARGNAME",
+"INDICATI","DRUGID","DRUGNAME",
+"HIGHEST_STATUS",
+"disease_name","disease_id","hpo_name",
+"CellType","ontLvl")] |>
+unique())
+#### Remove results that can't be linked to specific genes #####
+dat_sub <- ttdi$merged[!is.na(TARGETID) &
+!is.na(GENENAME2) &
+GENENAME2!="",]
+#### Filter by drug type ####
+if(!is.null(drug_types)){
+dat_sub <- dat_sub[
+grepl(paste(drug_types,collapse = "|"),DRUGNAME,ignore.case = TRUE) |
+grepl(paste(drug_types,collapse = "|"),DRUGTYPE,ignore.case = TRUE),]
+}
+#### Filter by status ####
+if(!is.null(keep_status)){
+dat_sub <- dat_sub[HIGHEST_STATUS %in% keep_status,]
+}
+if(!is.null(remove_status)){
+dat_sub <- dat_sub[!HIGHEST_STATUS %in% remove_status,]
+}
+dat_sub[,failed:=HIGHEST_STATUS %in% failed_status]
+#### Filter to only those in top_targets ####
+dat_sub2 <- (merge(
+dat_sub[failed==FALSE],
+top_targets,
+allow.cartesian = allow.cartesian,
+by.x = "GENENAME3",
+by.y = "gene_symbol")[,c("GENENAME2","TARGETID","TARGNAME",
+"INDICATI","DRUGID","DRUGNAME",
+"HIGHEST_STATUS",
+"disease_name","disease_id","hpo_name",
+"CellType","ontLvl")] |>
+unique())
+allow.cartesian
+allow.cartesian=T
+#### Filter to only those in top_targets ####
+dat_sub2 <- (merge(
+dat_sub[failed==FALSE],
+top_targets,
+allow.cartesian = allow.cartesian,
+by.x = "GENENAME3",
+by.y = "gene_symbol")[,c("GENENAME2","TARGETID","TARGNAME",
+"INDICATI","DRUGID","DRUGNAME",
+"HIGHEST_STATUS",
+"disease_name","disease_id","hpo_name",
+"CellType","ontLvl")] |>
+unique())
+#### Count proportion of drugs that our analyses captured ####
+pct_captured <- length(unique(dat_sub2$DRUGID)) /
+length(unique(dat_sub$DRUGID))*100
+# length(unique(paste0(dat_sub2$DRUGID,dat_sub2$INDICATI,
+#                      dat_sub2$GENENAME2)))
+dat_sub[,prioritised:=(DRUGID %in% dat_sub2$DRUGID)]
+#### Plot ####
+plt <- plot_ttd(dat_sub = dat_sub,
+failed_status = failed_status)
+#### Show ####
+if(isTRUE(show_plot)) methods::show(plt)
+#### Hypergeometric test ####
+stats::pyhypergeo::phypergeo(
+#### Hypergeometric test ####
+stats::phypergeo(
+k = length(unique(dat_sub2$DRUGID)),
+K = length(unique(dat_sub$DRUGID)),
+n = length(unique(dat_sub$DRUGID)),
+N = nrow(dat_sub)
+)
+#### Hypergeometric test ####
+stats::phyper(
+k = length(unique(dat_sub2$DRUGID)),
+K = length(unique(dat_sub$DRUGID)),
+n = length(unique(dat_sub$DRUGID)),
+N = nrow(dat_sub)
+)
+?stats::phyper
+dat_sub
+#### Hypergeometric test ####
+stats::phyper(
+q = length(unique(dat_sub2$DRUGID)),
+m = length(unique(dat_sub$DRUGID)),
+n = length(unique(dat_sub$DRUGID)),
+k = nrow(dat_sub)
+)
+#### Plot ####
+plt <- plot_ttd(dat_sub = dat_sub,
+failed_status = failed_status)
+is(plt)
+plt
+#### Hypergeometric test ####
+fail <- dat_sub[HIGHEST_STATUS %in% failed_status,drop=FALSE]
+notfail <- dat_sub[!HIGHEST_STATUS %in% failed_status,drop=FALSE]
+nrow(notfail)
+phyper(nrow(notfail)-1, nrow(notfail), (nrow(notfail)+nrow(top_targets))-nrow(notfail), nrow(top_targets),lower.tail= FALSE)
+notfail
+overlap <- nrow(notfail[prioritised==TRUE])
+group2 <- nrow(notfail)
+total <- nrow(notfail)+nrow(top_targets)
+group1 <- nrow(top_targets)
+## Test for over-representation (enrichment)
+# phyper(Overlap-1, group2, Total-group2, group1,lower.tail= FALSE)
+notfail <- dat_sub[!HIGHEST_STATUS %in% failed_status,drop=FALSE]
+overlap <- nrow(notfail[prioritised==TRUE])
+group2 <- nrow(notfail)
+total <- nrow(notfail)+nrow(top_targets)
+group1 <- nrow(top_targets)
+phyper(overlap-1,
+group2,
+total-group2,
+group1,
+lower.tail= FALSE)
+overlap
+overlap <- nrow(notfail[prioritised==TRUE])
+group2 <- nrow(notfail)
+total <- nrow(notfail)+nrow(notfail)
+group1 <- nrow(top_targets)
+total <- nrow(dat_sub)+nrow(top_targets)
+group1 <- nrow(top_targets)
+phyper(overlap-1,
+group2,
+total-group2,
+group1,
+lower.tail= FALSE)
+## Test for under-representation (depletion)
+phyper(overlap, group2, total-group2, group1, lower.tail= TRUE)
+overlap <- nrow(notfail[prioritised==TRUE])
+group2 <- nrow(notfail)
+total <- nrow(notfail)+nrow(top_targets)
+group1 <- nrow(top_targets)
+stats::phyper(overlap-1,
+group2,
+total-group2,
+group1,
+lower.tail= FALSE)
+## Test for under-representation (depletion)
+overlap <- nrow(fail[prioritised==TRUE])
+group2 <- nrow(fail)
+total <- nrow(fail)+nrow(top_targets)
+group1 <- nrow(top_targets)
+stats::phyper(overlap,
+group2,
+total-group2,
+group1,
+lower.tail= TRUE)
+total <- nrow(fail)#+nrow(top_targets)
+group1 <- nrow(top_targets)
+stats::phyper(overlap,
+group2,
+total-group2,
+group1,
+lower.tail= TRUE)
+group2 <- nrow(fail)
+total <- nrow(fail)+nrow(top_targets)
+group1 <- nrow(top_targets)
+stats::phyper(overlap,
+group2,
+total-group2,
+group1,
+lower.tail= TRUE)
+p2g <- HPOExplorer::load_phenotype_to_genes()
+unique(p2g$gene_symbol)
+length(unique(p2g$gene_symbol))
+## Test for over-representation (enrichment)
+# phyper(Overlap-1, group2, Total-group2, group1,lower.tail= FALSE)
+overlap <- nrow(notfail[prioritised==TRUE])
+group2 <- nrow(notfail)
+total <- length(unique(p2g$gene_symbol))
+group1 <- nrow(top_targets)
+nonfailed_enrichment <- stats::phyper(overlap-1,
+group2,
+total-group2,
+group1,
+lower.tail= FALSE)
+nonfailed_enrichment
+total
+overlap
+group2
+total
+str(notfail)
+## Test for over-representation (enrichment)
+# phyper(Overlap-1, group2, Total-group2, group1,lower.tail= FALSE)
+overlap <- data.table::uniqueN(notfail[prioritised==TRUE]$GENENAME3)
+group2 <- data.table::uniqueN(notfail$GENENAME3)
+total <- length(unique(p2g$gene_symbol))
+overlap
+group2
+total <- data.table::uniqueN(p2g$gene_symbol)
+group1 <- data.table::uniqueN(top_targets$gene_symbol)
+nonfailed_enrichment <- stats::phyper(overlap-1,
+group2,
+total-group2,
+group1,
+lower.tail= FALSE)
+nonfailed_enrichment
+group1
+total <- data.table::uniqueN(c(p2g$gene_symbol,notfail$GENENAME3))
+group1 <- data.table::uniqueN(top_targets$gene_symbol)
+nonfailed_enrichment <- stats::phyper(overlap-1,
+group2,
+total-group2,
+group1,
+lower.tail= FALSE)
+nonfailed_enrichment
+c(p2g$gene_symbol,notfail$GENENAME3)
+total <- data.table::uniqueN(c(p2g$gene_symbol,notfail$GENENAME3))
+group1 <- data.table::uniqueN(top_targets$gene_symbol)
+nonfailed_enrichment <- stats::phyper(overlap-1,
+group2,
+total-group2,
+group1,
+lower.tail= FALSE)
+nonfailed_enrichment
+#### Hypergeometric test ####
+dat_sub[,failed:=HIGHEST_STATUS %in% failed_status]
+fail <- dat_sub[failed==TRUE,drop=FALSE]
+notfail <- dat_sub[failed==TRUE,drop=FALSE]
+notfail
+## Test for under-representation (depletion)
+overlap <- data.table::uniqueN(fail[prioritised==TRUE]$GENENAME3)
+group2 <- data.table::uniqueN(fail$GENENAME3)
+total <- data.table::uniqueN(c(p2g$gene_symbol,fail$GENENAME3))
+group1 <- data.table::uniqueN(top_targets$gene_symbol)
+failed_depletion <- stats::phyper(overlap,
+group2,
+total-group2,
+group1,
+lower.tail= TRUE)
+failed_depletion
+nonfailed_enrichment
+source("~/Desktop/Rare Disease Celltyping/MSTExplorer/R/ttd_hypergeo.R")
+devtools::check_man()
+library(MSTExplorer)
+library(MSTExplorer)
+devoptera::args2vars(reassign = T)
+hpo_ids <- HPOExplorer::map_phenotypes(terms = phenotypes,
+to = "id")
+res <- data.table::copy(results)
+phenotypes <- c("Generalized neonatal hypotonia",
+"Scrotal hypospadias",
+"Increased circulating progesterone")
+# diseases_include <- "OMIM:176270"
+genes_include <- c("MAGEL2","HERC2")
+genes_exclude <- c("SNORD115-1")
+res
+#### Add diseases ####
+res <- HPOExplorer::add_disease(phenos = res,
+allow.cartesian = TRUE)
+res
+old_cols <- names(res)
+res <- add_driver_genes(res,
+phenotype_to_genes=phenotype_to_genes,
+metric = "specificity")
+#### Filter diseases ####
+if(!is.null(diseases_include)){
+res <- res[disease_id %in% diseases_include,]
+}
+if(!is.null(diseases_exclude)){
+res <- res[!disease_id %in% diseases_exclude,]
+}
+#### Add cell types ####
+res <- MSTExplorer::map_celltype(results = res)
+old_cols <- names(res)
+res <- add_driver_genes(res,
+phenotype_to_genes=phenotype_to_genes,
+metric = "specificity")
+ont=KGExplorer::get_ontology("uberon",method = "github")
+KGExplorer::filter_ontology(ont=ont, keep_descendants = "UBERON:0002038")
+ont=ontologyIndex::get_OBO("https://github.com/obophenotype/uberon/releases/download/v2024-03-22/uberon-base.obo")
+ont
+ont$children[["UBERON:0002038"]]
+ont$children
+ont$children["UBERON:0002038"]
+ont=ontologyIndex::get_OBO("https://github.com/obophenotype/uberon/releases/download/v2024-03-22/uberon-simple.obo")
+ont$children["UBERON:0002038"]
+ont=ontologyIndex::get_OBO("https://github.com/obophenotype/uberon/releases/download/v2024-03-22/uberon.obo")
+ont=ontologyIndex::get_OBO("https://github.com/obophenotype/uberon/releases/download/v2024-03-22/uberon.obo", propagate_relationships = T)
+ont
+ont$children["UBERON:0002038"]
+ont=ontologyIndex::get_OBO("https://github.com/obophenotype/uberon/releases/download/v2024-03-22/uberon.obo", merge_equivalent_terms = T)
+ont=simona::import_ontology("http://purl.obolibrary.org/obo/uberon/releases/2024-03-22/uberon.owl")
+ont=simona::import_ontology("http://purl.obolibrary.org/obo/uberon/releases/2024-03-22/uberon.owl")
+ont=simona::import_ontology("http://purl.obolibrary.org/obo/uberon/releases/2024-03-22/uberon.owl")
+ont@lt_children
+which(ont@terms=="UBERON:0002038")
+ont@lt_children[which(ont@terms=="UBERON:0002038")]
diff --git a/R/0docs.R b/R/0docs.R
index 2e60d80..0665294 100644
--- a/R/0docs.R
+++ b/R/0docs.R
@@ -19,6 +19,8 @@
 #' @param width Width of the saved plot.
 #' @param heights Passed to \link[patchwork]{wrap_plots}.
 #' @param subtitle_size Size of the plot subtitle.
+#' @param phenotype_to_genes Phenotype to gene mapping from
+#' \link[HPOExplorer]{load_phenotype_to_genes}.
 #' @family plot_
 #' @returns R object.
 #' @name plot_
diff --git a/R/plot_ontology_levels.R b/R/plot_ontology_levels.R
index f406f56..6e03fd7 100644
--- a/R/plot_ontology_levels.R
+++ b/R/plot_ontology_levels.R
@@ -13,6 +13,7 @@
 #' @param log_vars Logical vector indicating which variables to log-transform.
 #' @param sig_vars Logical vector indicating which variables to only plot
 #' for significant results.
+#' @param return_data Return the full long data used in the plots.
 #' @inheritParams plot_
 #' @inheritParams ggpubr::stat_cor
 #' @inheritParams prioritise_targets
@@ -57,7 +58,8 @@ plot_ontology_levels <- function(results = load_example_results(),
                          height=7,
                          width=length(x_vars)*5.75,
                          smooth.line.args=list(method = "loess",
-                                               se = FALSE)
+                                               se = FALSE),
+                         return_data=TRUE
                          ){
 
   requireNamespace("ggplot2")
@@ -252,6 +254,7 @@ plot_ontology_levels <- function(results = load_example_results(),
                           height = height,
                           width = width)
   }
+  if(isFALSE(return_data)) r2 <- NULL
   return(list(data=r2,
               data_stats=data_stats,
               plot=plts2))
diff --git a/R/plot_report.R b/R/plot_report.R
index 086d5ea..a2b8494 100644
--- a/R/plot_report.R
+++ b/R/plot_report.R
@@ -5,8 +5,6 @@
 #' @param rep_dt Report table.
 #' @param annot HPO annotations.
 #' @param remove_cols Columns to remove from \code{rep_dt}.
-#' @param phenotype_to_genes Phenotype to gene mapping from
-#' \link[HPOExplorer]{load_phenotype_to_genes}.
 #' @inheritParams plot_
 #' @inheritParams ggnetwork_plot_full
 #' @inheritDotParams ggplot2::ggsave
diff --git a/R/prioritise_targets.R b/R/prioritise_targets.R
index 45a8d05..990bb90 100644
--- a/R/prioritise_targets.R
+++ b/R/prioritise_targets.R
@@ -209,6 +209,10 @@ prioritise_targets <- function(#### Input data ####
     Severity_score <- cl_name <- cl_id <- Severity_score_max <-
     info_content <- NULL;
 
+  force(results)
+  force(ctd_list)
+  force(hpo)
+  force(phenotype_to_genes)
   t1 <- Sys.time()
   messager("Prioritising gene targets.",v=verbose)
   #### Add logFC ####
diff --git a/R/ttd_check.R b/R/ttd_check.R
index ef5333c..3135d09 100644
--- a/R/ttd_check.R
+++ b/R/ttd_check.R
@@ -34,18 +34,9 @@ ttd_check <- function(top_targets,
                       show_plot = TRUE,
                       save_path = NULL,
                       height=NULL,
-                      width=NULL){
-  # top_targets <- prioritise_targets(
-  #   keep_deaths = NULL,
-  #   keep_tiers = NULL,
-  #   severity_threshold_max = NULL,
-  #   severity_threshold = NULL,
-  #   pheno_frequency_threshold = NULL,
-  #   keep_onsets = NULL,
-  #   keep_ont_levels = seq(4),
-  #   keep_celltypes = NULL,
-  #   top_n = 2,
-  #   group_vars = c("hpo_id","disease_id"))$top_targets
+                      width=NULL,
+                      phenotype_to_genes=HPOExplorer::load_phenotype_to_genes()){
+  # top_targets <- prioritise_targets()$top_targets
   # drug_types <- c("Gene therapy"
     # "Antisense drug",
     # "Antisense oligonucleotide",
@@ -109,6 +100,14 @@ ttd_check <- function(top_targets,
   # length(unique(paste0(dat_sub2$DRUGID,dat_sub2$INDICATI,
   #                      dat_sub2$GENENAME2)))
   dat_sub[,prioritised:=(DRUGID %in% dat_sub2$DRUGID)]
+  #### Hypergeometric test ####
+  dat_sub[,failed:=HIGHEST_STATUS %in% failed_status]
+  fail <- dat_sub[failed==TRUE,drop=FALSE]
+  notfail <- dat_sub[failed==FALSE,drop=FALSE]
+  ttd_hypergeo_out <- ttd_hypergeo(fail=fail,
+                                   notfail=notfail,
+                                   top_targets=top_targets,
+                                   p2g=phenotype_to_genes)
   #### Plot ####
   plt <- plot_ttd(dat_sub = dat_sub,
                   failed_status = failed_status)
@@ -124,6 +123,7 @@ ttd_check <- function(top_targets,
     list(data=dat_sub,
          data_overlap=dat_sub2,
          pct_captured,
-         plot=plt)
+         plot=plt,
+         ttd_hypergeo_out=ttd_hypergeo_out)
   )
 }
diff --git a/R/ttd_hypergeo.R b/R/ttd_hypergeo.R
new file mode 100644
index 0000000..63b39b0
--- /dev/null
+++ b/R/ttd_hypergeo.R
@@ -0,0 +1,37 @@
+ttd_hypergeo <- function(fail,
+                         notfail,
+                         top_targets,
+                         p2g=HPOExplorer::load_phenotype_to_genes()){
+  # https://seqqc.wordpress.com/2019/07/25/how-to-use-phyper-in-r/
+
+  ## Test for over-representation (enrichment)
+  # phyper(Overlap-1, group2, Total-group2, group1,lower.tail= FALSE)
+  overlap <- data.table::uniqueN(notfail[prioritised==TRUE]$GENENAME3)
+  group2 <- data.table::uniqueN(notfail$GENENAME3)
+  total <- data.table::uniqueN(c(p2g$gene_symbol,notfail$GENENAME3))
+  group1 <- data.table::uniqueN(top_targets$gene_symbol)
+  nonfailed_enrichment <- stats::phyper(overlap-1,
+                                        group2,
+                                        total-group2,
+                                        group1,
+                                        lower.tail= FALSE)
+  messager("Non-failed gene targets enrichment p-value:",nonfailed_enrichment)
+  ## Test for under-representation (depletion)
+  # phyper(Overlap, group2, Total-group2, group1, lower.tail= TRUE)
+  overlap <- data.table::uniqueN(fail[prioritised==TRUE]$GENENAME3)
+  group2 <- data.table::uniqueN(fail$GENENAME3)
+  total <- data.table::uniqueN(c(p2g$gene_symbol,fail$GENENAME3))
+  group1 <- data.table::uniqueN(top_targets$gene_symbol)
+  failed_depletion <- stats::phyper(overlap,
+                                    group2,
+                                    total-group2,
+                                    group1,
+                                    lower.tail= TRUE)
+  messager("Failed gene targets depletion p-value:",failed_depletion)
+  return(
+    list(
+      nonfailed_enrichment=nonfailed_enrichment,
+      failed_depletion=failed_depletion
+    )
+  )
+}
diff --git a/man/plot_.Rd b/man/plot_.Rd
index d7877e2..1188dfa 100644
--- a/man/plot_.Rd
+++ b/man/plot_.Rd
@@ -36,6 +36,9 @@ Set to \code{NULL} to not save the plot.}
 \item{heights}{Passed to \link[patchwork]{wrap_plots}.}
 
 \item{subtitle_size}{Size of the plot subtitle.}
+
+\item{phenotype_to_genes}{Phenotype to gene mapping from
+\link[HPOExplorer]{load_phenotype_to_genes}.}
 }
 \value{
 R object.
diff --git a/man/predict_celltypes.Rd b/man/predict_celltypes.Rd
index 061ad46..6e90a03 100644
--- a/man/predict_celltypes.Rd
+++ b/man/predict_celltypes.Rd
@@ -51,9 +51,8 @@ genes in the include, default, and exclude lists.}
 \link[MSTExplorer]{gen_results}
 and merged together with \link[MSTExplorer]{merge_results}}
 
-\item{phenotype_to_genes}{Output of
-\link[HPOExplorer]{load_phenotype_to_genes} mapping phenotypes
-to gene annotations.}
+\item{phenotype_to_genes}{Phenotype to gene mapping from
+\link[HPOExplorer]{load_phenotype_to_genes}.}
 
 \item{agg_var}{The variable(s) to aggregate \code{results} by.}
 
diff --git a/man/ttd_check.Rd b/man/ttd_check.Rd
index 22c8cca..f372fd3 100644
--- a/man/ttd_check.Rd
+++ b/man/ttd_check.Rd
@@ -14,7 +14,8 @@ ttd_check(
   show_plot = TRUE,
   save_path = NULL,
   height = NULL,
-  width = NULL
+  width = NULL,
+  phenotype_to_genes = HPOExplorer::load_phenotype_to_genes()
 )
 }
 \arguments{
@@ -40,6 +41,9 @@ Set to \code{NULL} to not save the plot.}
 \item{height}{Height of the saved plot.}
 
 \item{width}{Width of the saved plot.}
+
+\item{phenotype_to_genes}{Phenotype to gene mapping from
+\link[HPOExplorer]{load_phenotype_to_genes}.}
 }
 \description{
 Identify the overlap between your prioritised list of gene therapy targets