use filters arg to be more flexible

DESCRIPTION

@@ -29,7 +29,7 @@ RoxygenNote: 7.3.1
 Depends: 
     R (>= 4.1)
 Imports:
-    HPOExplorer (>= 0.99.10),
+    HPOExplorer (>= 1.0.0),
     KGExplorer,
     orthogene,
     data.table,
@@ -61,7 +61,8 @@ Suggests:
     patchwork,
     ggpubr, 
     igraph,
-    orthogene
+    ggdendro,
+    dendextend
 Remotes: 
   github::neurogenomics/HPOExplorer,
   github::neurogenomics/KGExplorer,

NAMESPACE

@@ -5,8 +5,6 @@ export(agg_results)
 export(correlation_heatmap)
 export(create_dt)
 export(ewce_para)
-export(frequency_barplot)
-export(frequency_histogram)
 export(gen_overlap)
 export(gen_results)
 export(get_bg)
@@ -18,14 +16,16 @@ export(map_celltype)
 export(map_tissue)
 export(map_tissues)
 export(merge_results)
-export(multiewce_plot)
+export(plot_bar_dendro)
+export(plot_bar_summary)
+export(plot_frequency_bar)
+export(plot_frequency_histogram)
 export(plot_ont_lvl)
+export(plot_report)
 export(prioritise_targets)
 export(prioritise_targets_network)
-export(report_plot)
 export(subset_phenos)
 export(subset_results)
-export(summary_plot)
 export(terminal_celltypes)
 export(ttd_check)
 import(GeneOverlap)
@@ -34,10 +34,8 @@ import(KGExplorer)
 import(data.table)
 import(ggplot2)
 importFrom(EWCE,bootstrap_enrichment_test)
-importFrom(EWCE,ewce_plot)
 importFrom(HPOExplorer,add_ancestor)
 importFrom(HPOExplorer,add_gene_frequency)
-importFrom(HPOExplorer,filter_descendants)
 importFrom(HPOExplorer,get_gene_lists)
 importFrom(HPOExplorer,load_phenotype_to_genes)
 importFrom(data.table,":=")

NEWS.md

@@ -160,10 +160,10 @@
   - `prioritise_targets`
   - `prioritise_targets_heatmap`
   - `prioritise_targets_network`
-  - `report_plot`
+  - `plot_report`
   - `correlation_heatmap`
   - `frequency_bar`
-  - `frequency_histogram` 
+  - `plot_frequency_histogram` 
 
 * Added `example_targets` data:
   - Speeds up run time of examples/tests.

R/add_ctd.R

@@ -9,7 +9,7 @@
 #' \item{"mean_exp_quantiles"} Mean expression quantile per cell-type.
 #' }
 #' @inheritParams prioritise_targets
-#' @inheritParams report_plot
+#' @inheritParams plot_report
 #' @inheritParams data.table::merge.data.table
 #'
 #' @export

R/agg_results.R

@@ -3,7 +3,7 @@
 #' Aggregate enrichment results from generated by \link[MultiEWCE]{gen_results}.
 #' Counts unique entries in \code{count_var} while grouping by \code{group_var}.
 #' @param sep Separator for collapsed character columns.
-#' @inheritParams summary_plot
+#' @inheritParams plot_bar_summary
 #' @inheritParams HPOExplorer::make_phenos_dataframe
 #' @inheritParams HPOExplorer::make_network_object
 #' @returns Aggregated \link[data.table]{data.table}
@@ -13,7 +13,7 @@
 #' @importFrom stringr str_wrap
 #' @import HPOExplorer
 #' @examples
-#' phenos <- subset_results(cell_type = "Microglia")
+#' phenos <- subset_results(filters = list(CellType="Microglia"))
 #' agg_res <- agg_results(phenos = phenos)
 agg_results <- function(phenos,
                         count_var = "hpo_name",

R/ggnetwork_plot_full.R

@@ -5,21 +5,15 @@
 #'
 #' @param results The cell type-phenotype enrichment results generated by
 #'  \link[MultiEWCE]{gen_results}
-#'  and merged together with \link[MultiEWCE]{merge_results}.
-#' @param cell_type The cell type of interest to be plotted.
-#' Can be a single string (e.g. \code{"Astrocytes"}) or a character vector
-#' of multiple cell types (e..g. \code{c("Astrocytes","Microglia")}).
-#' Set to \code{NULL} if you wish to include all cell-types that are available
-#' (after \code{q_threshold} and \code{fold_threshold} filtering).
-#' If >1 cell-type remains, results will be aggregated automatically
-#' such that there is only 1 node per phenotype.
+#'  and merged together with \link[MultiEWCE]{merge_results}
 #' @param q_threshold The q value threshold to subset the \code{results} by.
 #' @param fold_threshold The minimum fold change in specific expression
 #'  to subset the \code{results} by.
 #' @inheritParams HPOExplorer::make_phenos_dataframe
 #' @inheritParams HPOExplorer::make_network_plot
 #' @inheritParams HPOExplorer::filter_descendants
 #' @inheritParams HPOExplorer::make_network_object
+#' @inheritParams KGExplorer::filter_dt
 #' @inheritDotParams HPOExplorer::make_network_plot
 #' @returns A named list of outputs,
 #'  including a interactive network plot of the selected subset
@@ -29,8 +23,8 @@
 #' @import HPOExplorer
 #' @import KGExplorer
 #' @examples
-#' res <- ggnetwork_plot_full(cell_type = "Microglia")
-ggnetwork_plot_full <- function(cell_type,
+#' res <- ggnetwork_plot_full(filters = list(CellType = "Microglia"))
+ggnetwork_plot_full <- function(filters,
                                 keep_descendants = NULL,
                                 results = load_example_results(),
                                 hpo = HPOExplorer::get_hpo(),
@@ -46,16 +40,15 @@ ggnetwork_plot_full <- function(cell_type,
                                 verbose = TRUE,
                                 ...){
   messager("ggnetwork_plot_full",v=verbose)
-  phenos <- subset_phenos(keep_descendants = keep_descendants,
+  phenos <- subset_phenos(filters = filters,
+                          keep_descendants = keep_descendants,
                           results = results,
                           hpo = hpo,
-                          cell_type = cell_type,
                           q_threshold = q_threshold,
                           fold_threshold = fold_threshold,
                           verbose = verbose)
-  phenos$hpo_name <- HPOExplorer::map_phenotypes(terms = phenos$hpo_id,
-                                                 hpo = hpo,
-                                                 to="name")
+  phenos <- HPOExplorer::add_hpo_name(phenos = phenos,
+                                      hpo = hpo)
   #### Aggregate across multiple celltypes ####
   phenos <- agg_results(phenos = phenos,
                         count_var = "CellType",

R/map_celltype.R

@@ -6,6 +6,7 @@
 #' @param rm_prefixes Prefixes to remove from cell type names in
 #' \code{input_col} before performing mapping.
 #' @param by Columns to merge on.
+#' @param input_col Column to use for linking with the \code{map} data.
 #' @inheritParams ggnetwork_plot_full
 #' @export
 #' @import data.table
@@ -25,6 +26,7 @@ map_celltype <- function(results,
   if(all(new_cols %in% names(results))) {
     return(results)
   }
+  messager("Mapping cell types to cell ontology terms.")
   results[,author_celltype:=gsub(paste(paste0("^",rm_prefixes,"_"),
                                        collapse = "|"),"",
                                  get(input_col),ignore.case = TRUE)]

R/map_tissue.R

@@ -14,11 +14,14 @@ map_tissue <- function(results,
                            package = "MultiEWCE",
                            name="anatomy_maps")
                        ){
+  ancestor <- ancestor_name <- tissue_ontology_term_id <- tissue <- n_tissues <-
+    n_ancestors <- cl_name <- NULL;
   results <- map_celltype(results)
   new_cols <- c("tissue_ontology_term_id","tissue")
   if(all(new_cols %in% names(results))) {
     return(results)
   }
+  messager("Mapping cell types to tissue ontology terms.")
   cl <- KGExplorer::get_ontology("cl")
   #### Assign each tissue an ancestral group ####
   uberon <- KGExplorer::get_ontology("uberon",
@@ -50,15 +53,15 @@ map_tissue <- function(results,
   #### Create tissue-celltype heatmap #####
 
   ggplot2::ggplot(map2,
-                  ggplot2::aes(x=cell_type, y=ancestor_name,
+                  ggplot2::aes(x=cl_name, y=ancestor_name,
                                fill=ancestor_name)
                   ) +
     ggplot2::geom_tile(show.legend = FALSE) +
     ggplot2::theme(axis.text.x = ggplot2::element_text(angle = 90,
                                                        hjust = 1, vjust=0.5))
 
 
-  return(results_cl)
+  return(map2)
 }
 
 

R/multiewce_plot.R → R/plot_bar_dendro.R

@@ -1,28 +1,38 @@
-#' EWCE plot
+#' Plot bar dendrogram
 #'
-#' A shallow wrapper for the \link[EWCE]{ewce_plot} function.
-#' @inheritParams EWCE::ewce_plot
-#' @inheritParams HPOExplorer::add_ancestor
+#' Create a plot summarising MultiEWCE results as a bar chart with multiple
+#'  facets and a cell ontology-based dendrogram.
+#' @param ancestor_names Keep terms with specific ancestors.
+#' @param celltype_col Name of the cell type column in the \code{results}.
 #' @inheritParams ggnetwork_plot_full
+#' @inheritParams HPOExplorer::add_ancestor
+#' @inheritParams ggplot2::facet_wrap
 #' @inheritDotParams EWCE::ewce_plot
 #' @returns A bar chart with dendrogram of EWCE results in each cell type.
 #'
 #' @export
-#' @importFrom EWCE ewce_plot
 #' @examples
 #' results <- load_example_results(multi_dataset=TRUE)
-#' out <- multiewce_plot(results = results)
-multiewce_plot <- function(results = load_example_results(multi_dataset = TRUE),
-                           ancestor_names=c(
-                             "Abnormality of the nervous system",
-                             "Abnormality of the cardiovascular system",
-                             "Abnormality of the immune system",
-                             "Abnormality of the respiratory system",
-                             "Abnormality of the eye",
-                             "Abnormality of the endocrine system",
-                             "Neoplasm"),
-                      ...) {
+#' out <- plot_bar_dendro(results = results)
+plot_bar_dendro <- function(results = load_example_results(multi_dataset = TRUE),
+                            celltype_col = "cl_name",
+                            ancestor_names=c(
+                              "Abnormality of the nervous system",
+                              "Abnormality of the cardiovascular system",
+                              "Abnormality of the immune system",
+                              "Abnormality of the respiratory system",
+                              "Abnormality of the eye",
+                              "Abnormality of the endocrine system",
+                              "Neoplasm"),
+                            facets = "ancestor_name",
+                            ...) {
   requireNamespace("ggplot2")
+  requireNamespace("patchwork")
+  requireNamespace("dendextend")
+  requireNamespace("ggdendro")
+  hpo_id <- p <- fold_change <- cl_name <- ancestor_name <-
+    enriched_phenotypes_norm <- enriched_phenotypes <- top_ancestor_name <-
+    color <- NULL;
 
   results <- map_celltype(results)
   #### Get celltype dendrogram ####
@@ -35,27 +45,29 @@ multiewce_plot <- function(results = load_example_results(multi_dataset = TRUE),
   #### Method 1 ####
   # X <- KGExplorer::ontology_to(ont = cl,
   #                              to="similarity")
-  # X <- X[unique(results$cell_type_ontology_term_id),
-  #        unique(results$cell_type_ontology_term_id)]
+  # X <- X[unique(results$cl_id),
+  #        unique(results$cl_id)]
   # hc <- stats::hclust(stats::dist(X) )
   # ddata <- ggdendro::dendro_data(hc)
   #### Method 2 ####
   # g <- KGExplorer::ontology_to(ont = cl,
   #                               to="tbl_graph")
-  # KGExplorer::filter_graph(g, node_filters = list(name=results$cell_type_ontology_term_id))
+  # KGExplorer::filter_graph(g,
+  # node_filters = list(name=results$cl_id))
 
   #### Method 3 ####
   hc <- KGExplorer::ontology_to(ont = cl,
                                 to="igraph_dist_hclust")
-  dend <- as.dendrogram(hc)
-  dend2 <- dendextend::prune(dend,
-                             leaves=labels(dend)[!labels(dend) %in% results$cell_type_ontology_term_id]
-                             )
+  dend <- stats::as.dendrogram(hc)
+  messager("Pruning dendrogram.")
+  dend2 <- dendextend::prune(
+    dend,
+    leaves=labels(dend)[!labels(dend) %in% results$cl_id])
   ddata <- ggdendro::dendro_data(dend2)
 
 
   ddata$labels$label <- KGExplorer::map_ontology_terms(ont = cl,
-                                                      terms=ddata$labels$label)
+                                                       terms=ddata$labels$label)
   ggdend <-
     ggdendro::ggdendrogram(ddata) +
     ggplot2::scale_x_discrete(expand =c(.01, .01)) +
@@ -82,20 +94,21 @@ multiewce_plot <- function(results = load_example_results(multi_dataset = TRUE),
                               p=mean(p),
                               q=mean(q),
                               fold_change=mean(fold_change)),
-                 by=c("CellType","cell_type","cell_type_ontology_term_id",
-                      "ancestor","ancestor_name")]
-  dat <- dat[cell_type %in% unique(ddata$labels$label),]
-  dat <- dat[ancestor_name %in% ancestor_names,]
+                 by=c(celltype_col,"ancestor","ancestor_name")]
+
+  dat <- dat[get(celltype_col) %in% unique(ddata$labels$label),]
+  dat <- KGExplorer::filter_dt(dat,
+                               filters = list(ancestor_name=ancestor_names))
 
-  dat$cell_type <- factor(dat$cell_type,
-                          levels=unique(ddata$labels$label),
-                          ordered = TRUE)
+  dat[[celltype_col]] <- factor(dat[[celltype_col]],
+                                levels=unique(ddata$labels$label),
+                                ordered = TRUE)
   #### Get the top HPO category for each cell type ####
   ### NOrmalise count within each ancestor
-  dat[,enriched_phenotypes_norm:=enriched_phenotypes/max(enriched_phenotypes),
+  dat[,enriched_phenotypes_norm:=(enriched_phenotypes/max(enriched_phenotypes)),
       by=c("ancestor")]
   dat[,top_ancestor_name:=head(ancestor_name[which.max(enriched_phenotypes_norm)],1),
-      by=c("cell_type","cell_type_ontology_term_id")]
+      by=c(celltype_col)]
   # cellmeta <- data.table::data.table(
   #   cl@elementMetadata[,c("name","ancestor","ancestor_name")],
   #   key = "name"
@@ -105,36 +118,38 @@ multiewce_plot <- function(results = load_example_results(multi_dataset = TRUE),
                                          columns = "top_ancestor_name",
                                          preferred_palettes = "brewer.set2",
                                          as="dict")[[1]]
-  dat2 <- unique(dat[,c("cell_type","top_ancestor_name")]
+  dat2 <- unique(dat[,c(celltype_col,"top_ancestor_name"), with=FALSE]
                  )[,color:=color_dict[top_ancestor_name]] |>
-    data.table::setkeyv("cell_type")
+    data.table::setkeyv(celltype_col)
   color_vector <- dat2[unique(ddata$labels$label),]$color
   color_vector[is.na(color_vector)] <- "grey20"
-  # color_vector <- KGExplorer::map_colors(unique(dat[,c("cell_type","top_ancestor_name")]),
+  # color_vector <- KGExplorer::map_colors(unique(dat[,c(celltype_col,"top_ancestor_name")]),
   #                                        columns = "top_ancestor_name",
   #                                        preferred_palettes = "brewer.set2",
   #                                      as="vector")[[1]]
 
-  ggbars <- ggplot2::ggplot(dat, ggplot2::aes(x=cell_type,
+  ggbars <- ggplot2::ggplot(dat, ggplot2::aes(x=!!ggplot2::sym(celltype_col),
                                               y=enriched_phenotypes,
                                               fill=ancestor_name)) +
     ggplot2::geom_bar(stat="identity", show.legend = FALSE) +
     ggplot2::scale_fill_manual(values=color_dict) +
-    ggplot2::facet_wrap(facets = "ancestor_name",
+    ggplot2::facet_wrap(facets = facets,
                         scales="free_y", ncol = 1) +
     ggplot2::labs(x=NULL,y="Enriched phenotypes") +
     ggplot2::theme_bw() +
     ggplot2::theme(axis.text.x = ggplot2::element_text(
       angle = 90, hjust = 1, vjust=0.5,
       color = unname(color_vector)
       ),
-          strip.background = ggplot2::element_rect(fill = "transparent")
+      strip.background = ggplot2::element_rect(fill = "transparent")
+      )
 
-          )
-
-  patchwork::wrap_plots(ggbars ,
+  ggp <- patchwork::wrap_plots(ggbars ,
                         ggdend + ggplot2::scale_y_reverse(),
-
                         ncol = 1, heights = c(1,.3))
-
+  #### Return ####
+  return(
+    list(data=dat,
+         plot=ggp)
+  )
 }