Merge branch 'master' into docs-readme

.github/workflows/rworkflows.yml

@@ -38,7 +38,7 @@ jobs:
           cont: ~
           rspm: ~
     steps:
-    - uses: neurogenomics/rworkflows@dev
+    - uses: neurogenomics/rworkflows@master
       with:
         run_bioccheck: ${{ false }}
         run_rcmdcheck: ${{ true }}

DESCRIPTION

@@ -1,6 +1,6 @@
 Package: MSTExplorer
 Title: Multi-Scale Target Explorer
-Version: 1.0.4
+Version: 1.0.6
 Authors@R:
     c(
     person(given = "Brian",
@@ -54,7 +54,8 @@ Imports:
     future,
     furrr,
     tidygraph,
-    rstatix
+    rstatix,
+    simona
 Suggests: 
     rmarkdown,
     knitr,

NEWS.md

@@ -1,3 +1,52 @@
+# MSTExplorer 1.0.6
+
+## Bug fixes
+* `get_color_map`
+  - Fix `color_vector` assignment.
+
+# MSTExplorer 1.0.5
+
+## Bug fixes
+
+* Tests
+  - Use `force_new=TRUE` where some tests occasionally fail with cached
+  files.
+  - `test-load_example_results`: Update test files.
+  - `test-prioritise_targets`: Remove unused arguments and change input size.
+  - `test-prioritise_targets_network`: Process `top_targets` to include effect
+  - `test-plot_differential_outcomes`: Use non-specific plot name in
+  `patchwork::wrap_plots`.
+  - `test-plot_differential_outcomes`: Wrap p3 in `expect_error` to prevent
+  test failure even if error was handled.
+  - `test-report_plot`: Fix detection for ggplot object.
+  variable.
+* Vignettes
+  - `MSTExplorer`: Update effect variable to `fold_change`.
+* `add_logfc`
+  - Return `results` with new column rather than directly modifying the original
+  input.
+  - Update references (`add_logfc(results)` -> `results <- add_logfc(results)`).
+* `ttd_check`, `plot_differential_outcomes`, `prioritise_targets_grid`
+  - Add check for `disease_name` column before executing
+  `HPOExplorer::add_disease` on input.
+* `plot_ttd`
+  - Remove `fill` aesthetic for `geom_text` (doesn't exist anymore).
+* `extract_help`
+  - [DEVELOPEMENT ONLY] Look for help docs only in legitimate pkg installation
+  paths.
+* `subset_results`
+  - Add new `effect_var` argument.
+  - Adjust default `effect_threshold` to 0.1.
+* `add_symptom_results`
+  - Only merge `results` and `phenotypes_to_genes` if required (prevents
+  column duplicates with altered names).
+* `map_tissue`
+  - Rewrite merge logic to fix error: attempt to replicate non-vector.
+* `test_target_celltypes`
+  - Uncomment `add_ancestors(results)` to ensure required ancestor columns are
+  present.
+* Add missing import: simona
+
 # MSTExplorer 1.0.4
 
 ## New features

R/add_logfc.R

@@ -14,13 +14,12 @@ add_logfc <- function(results,
   logFC <- NULL;
   if(!"logFC" %in% names(results) || isTRUE(force_new)){
     messager("Adding logFC column.")
-    # results[,logFC:=log10(scales::rescale(estimate,c(.Machine$double.xmin, 1)))]
-    # results[,logFC:=logFC/mean(results$logFC)]
-    results[,logFC:=(get(effect_var)+abs(min(get(effect_var))))]
-    results[,logFC:=(log2(logFC/mean(results$logFC)))]
-    # hist(results$logFC)
+    results$logFC <- results[[effect_var]]+abs(min(results[[effect_var]]))
+    results$logFC <- log2(results$logFC/mean(results$logFC))
+    return(results)
   } else {
     messager("logFC already exists in results.",
              "Use `force_new=TRUE` to overwrite.")
+    return(results)
   }
 }

R/add_symptom_results.R

@@ -16,8 +16,8 @@
 #'
 #' @export
 #' @examples
-#' results = load_example_results()[seq(5000)]
-#' results <- add_symptom_results()
+#' results <- load_example_results()[seq(5000)]
+#' results <- add_symptom_results(results)
 add_symptom_results <- function(results = load_example_results(),
                                 q_threshold = 0.05,
                                 effect_threshold = NULL,
@@ -57,13 +57,19 @@ add_symptom_results <- function(results = load_example_results(),
   results <- HPOExplorer::add_genes(results,
                                     phenotype_to_genes = phenotype_to_genes,
                                     allow.cartesian = TRUE)
-  results_annot <- data.table::merge.data.table(
-    results,
-    unique(phenotype_to_genes[,c("hpo_id","disease_id",
-                                 "n_genes_hpo_id",
-                                 "n_genes_disease_id",
-                                 "n_genes_symptom")]),
-    by=c("hpo_id","disease_id"))
+  merge_cols <- c("hpo_id",
+                  "disease_id",
+                  "n_genes_hpo_id",
+                  "n_genes_disease_id",
+                  "n_genes_symptom")
+  if(all(merge_cols %in% colnames(results))){
+    results_annot <- results
+  }else{
+    results_annot <- data.table::merge.data.table(
+      results,
+      unique(phenotype_to_genes[,..merge_cols]),
+      by=c("hpo_id","disease_id"))
+  }
   #### Add genes that intersect between the
   phenos <- add_driver_genes(results = results_annot,
                              ctd_list = ctd_list,

R/extract_help.R

@@ -14,7 +14,7 @@ extract_help <- function(pkg,
   requireNamespace("tools")
 
   to <- match.arg(to)
-  rdbfile <- file.path(find.package(pkg), "help", pkg)
+  rdbfile <- file.path(find.package(pkg, lib.loc = .libPaths()), "help", pkg)
   fetchRdDB <- utils::getFromNamespace("fetchRdDB","tools")
   rdb <- fetchRdDB(rdbfile, key = fn)
   convertor <- switch(to,

R/get_color_map.R

@@ -14,7 +14,7 @@ get_color_map <- function(dat,
   if(!is.null(celltype_col_order)){
     dat2 <- dat2[order(match(dat2[[celltype_col]],celltype_col_order)),]
   }
-  dat2[,ancestor_color:=color_map[top_ancestor_name]]
+  dat2$ancestor_color<-color_map[dat2[[columns]]]
   color_vector <- dat2$ancestor_color
   return(list(
     color_map=color_map,

R/map_tissue.R

@@ -47,12 +47,15 @@ map_tissue <- function(results = NULL,
     by=by]
   #### Get the ancestor for each tissue #####
   if(!isFALSE(lvl)){
+    ancestor_dat <- data.frame(uberon@elementMetadata@listData) |>
+      dplyr::select(id,ancestor,ancestor_name) |>
+      dplyr::rename(
+        uberon_id = id,
+        uberon_ancestor = ancestor,
+        uberon_ancestor_name = ancestor_name
+      )
     map_agg2 <- merge(map_agg,
-                      data.table::data.table(
-                        uberon@elementMetadata
-                      )[,list(uberon_id=id,
-                              uberon_ancestor=ancestor,
-                              uberon_ancestor_name=ancestor_name)],
+                      ancestor_dat,
                       all.x=TRUE,
                       by.x="top_uberon_id",
                       by.y="uberon_id")

R/plot_bar_dendro.R

@@ -87,7 +87,7 @@ plot_bar_dendro <- function(results = load_example_results(),
   }
   #### Filter the results ####
   by <- unique(c(celltype_col,"cl_id","ancestor","ancestor_name"))
-  add_logfc(results)
+  results <- add_logfc(results)
   dat <- results[q<q_threshold & cl_id %in% unique(ddata$labels$id),
                  lapply(.SD,mean),
                  by=by,
@@ -104,7 +104,7 @@ plot_bar_dendro <- function(results = load_example_results(),
                 normalise_group=TRUE)
   #### Color each cell type x-axis label by the most commonly enriched HPO category ####
   cmap <- get_color_map(dat=dat,
-                        columns = "top_ancestor_name",
+                        columns = "ancestor_name",
                         ddata=ddata,
                         celltype_col=celltype_col,
                         preferred_palettes=preferred_palettes)

R/plot_bar_dendro_facets.R

@@ -168,7 +168,7 @@ plot_bar_dendro_facets <- function(results = load_example_results(),
       data_summary <- NULL
     } else {
       cl_name <- p.value.adj <- term <- p.value.adj.signif <- NULL;
-      dat <- merge(dat[,-c("term")],
+      dat <- merge(dat,
                    target_tests[[1]][term=="is_targetTRUE"],
                    all.x = TRUE,
                    by=c("ancestor_name","cl_id"))

R/plot_differential_outcomes.R

@@ -39,6 +39,10 @@ plot_differential_outcomes <- function(results,
 
   {
     results <- HPOExplorer::add_hpo_name(results)
+    if ("disease_name" %in% c(facet_var, x_var, y_var) &&
+        !"disease_name" %in% names(results)) {
+      results <- HPOExplorer::add_disease(results, add_descriptions = TRUE)
+    }
     plot_dat <- results[!is.na(get(y_var)),]
     plot_dat[,n_celltype:=length(unique(get(x_var))),by=c(facet_var)]
     plot_dat[,n_ids:=.N, by=c(facet_var)]

R/plot_ttd.R

@@ -33,7 +33,7 @@ plot_ttd <- function(dat_sub,
         size=label_size,
         angle=90,
         hjust=1,
-        ggplot2::aes(x=HIGHEST_STATUS, fill=prioritised,
+        ggplot2::aes(x=HIGHEST_STATUS,
                      label=paste0("n=",n_drugs), y=.95),
         # fill="white", alpha=.75,
         inherit.aes = FALSE) +

R/predict_celltypes.R

@@ -78,7 +78,7 @@ predict_celltypes <- function(phenotypes,
     hpo_ids <- HPOExplorer::map_phenotypes(terms = phenotypes,
                                            to = "id")
     res <- data.table::copy(results)
-    add_logfc(res)
+    res <- add_logfc(res)
     res[,effect:=scales::rescale(get(effect_var),c(0,1))]
     # res <- res[q<0.05]
     # res[,effect:=scales::rescale(abs(effect))]

R/prioritise_targets.R

@@ -265,7 +265,7 @@ prioritise_targets <- function(#### Input data ####
   t1 <- Sys.time()
   messager("Prioritising gene targets.",v=verbose)
   #### Add logFC ####
-  add_logfc(results = results)
+  results <- add_logfc(results)
   #### add_hpo_id  #####
   results <- HPOExplorer::add_hpo_id(phenos = results,
                                      hpo = hpo)

R/prioritise_targets_grid.R

@@ -85,10 +85,14 @@ prioritise_targets_grid <- function(top_targets,
                         size=3){
     if(!is.null(labels)) labels <- stats::setNames(labels,as.character(x_pos))
     dat <- merge(dt1[,-c("hpo_id")],
-                 dt2, by="hpo_name", sort=FALSE)|>
-      data.table::melt.data.table(
-        id.vars=c("hpo_id","hpo_name","severity_class","p"),
-        measure.vars=unique(c(label_cols,"gene_symbol")))
+                 dt2, by="hpo_name", sort=FALSE)
+    # Check for disease_name
+    if("disease_name" %in% label_cols && !"disease_name" %in% names(dat)){
+      dat <- HPOExplorer::add_disease(dat, add_descriptions = TRUE)
+    }
+    dat <- data.table::melt.data.table(dat,
+            id.vars=c("hpo_id","hpo_name","severity_class","p"),
+            measure.vars=unique(c(label_cols,"gene_symbol")))
     dat[,hpo_name:=factor(hpo_name,
                           levels=rev(levels(dt1$hpo_name)), ordered=TRUE)]
     dat[startsWith(value,"GRANULOMATOUS"), value:=stringr::str_to_title(value)]

R/prioritise_targets_network.R

@@ -88,7 +88,7 @@ prioritise_targets_network <- function(top_targets,
   requireNamespace("ggplot2")
   requireNamespace("pals")
 
-  add_logfc(top_targets)
+  top_targets <- add_logfc(top_targets)
   if(any(c("cl_name","cl_id") %in% vertex_vars)){
     map_celltype(top_targets)
   }

R/subset_results.R

@@ -4,6 +4,7 @@
 #' @inheritParams ggnetwork_plot_full
 #' @inheritParams KGExplorer::filter_dt
 #' @inheritParams HPOExplorer::make_phenos_dataframe
+#' @inheritParams add_logfc
 #' @returns A data frame of the selected subset of RD EWCE results
 #' with HPO ID column added.
 #'
@@ -13,7 +14,8 @@
 subset_results <- function(filters = list(cl_name=NULL),
                            results = load_example_results(),
                            q_threshold = 0.0005,
-                           effect_threshold = 1,
+                           effect_threshold = 0.1,
+                           effect_var = "fold_change",
                            verbose = TRUE){
   effect <-  hpo_id <- hpo_id <- NULL;
 
@@ -24,7 +26,7 @@ subset_results <- function(filters = list(cl_name=NULL),
     results_sig <- results_sig[q<q_threshold]
   }
   if(is.numeric(effect_threshold)){
-    results_sig <- results_sig[effect>effect_threshold]
+    results_sig <- results_sig[results_sig[[effect_var]]>effect_threshold,]
   }
   #### Check that celltype is available ####
   if(nrow(results_sig)==0){

R/test_target_celltypes.R

@@ -40,7 +40,7 @@ test_target_celltypes <- function(results=load_example_results(),
   ancestor_name <- is_sig <- is_target <- valid <- cl_id <- NULL;
 
   method <- match.arg(method)
-  # results <- HPOExplorer::add_ancestor(results)
+  results <- HPOExplorer::add_ancestor(results)
   results <- map_celltype(results)
   results[,is_sig:=q<q_threshold][,is_target:=get(celltype_col) %in%
                                   target_celltypes[[get(ancestor_var)]],

R/ttd_check.R

@@ -67,8 +67,12 @@ ttd_check <- function(top_targets,
   TARGETID <- DRUGNAME <- DRUGTYPE <- DRUGID <-
     GENENAME2 <- prioritised <- HIGHEST_STATUS <- NULL;
 
-  top_targets <- HPOExplorer::add_disease(top_targets,
+  #### add_disease if not already done ###
+  if (!"disease_name" %in% names(top_targets) |
+      !"disease_id" %in% names(top_targets)) {
+    top_targets <- HPOExplorer::add_disease(top_targets,
                                           add_descriptions = TRUE)
+  }
   ttdi <- KGExplorer::get_ttd(force_new = force_new,
                               run_map_genes = run_map_genes)
   #### Remove results that can't be linked to specific genes #####

R/validate_associations_correlate_ctd.R

@@ -29,7 +29,7 @@ validate_associations_correlate_ctd <- function(results=load_example_results(),
                                                 ...){
   test_id <- NULL;
   results <- map_celltype(results)
-  add_logfc(results)
+  results <- add_logfc(results)
   results <- KGExplorer::filter_dt(results,
                                    filters = filters)
   group_values <- unique(results[[group_var]])

R/validate_associations_mkg.R

@@ -30,7 +30,7 @@ validate_associations_mkg <- function(results=load_example_results(),
                                       ){
   from <- hpo_id <- NULL;
   # kg <- data.table::fread(here::here("data/monarch_kg_cells.csv"))
-  add_logfc(results)
+  results <- add_logfc(results)
   kg_og <- data.table::copy(kg)
   kg <- kg[grepl("HP:",from)][from %in% unique(results$hpo_id)]
   message(paste(

README.md

@@ -3,7 +3,7 @@ MSTExplorer
 <img src='https://github.com/neurogenomics/MSTExplorer/raw/master/inst/hex/hex.png' title='Hex sticker for MSTExplorer' height='300'><br>
 [![License:
 GPL-3](https://img.shields.io/badge/license-GPL--3-blue.svg)](https://cran.r-project.org/web/licenses/GPL-3)
-[![](https://img.shields.io/badge/devel%20version-1.0.4-black.svg)](https://github.com/neurogenomics/MSTExplorer)
+[![](https://img.shields.io/badge/devel%20version-1.0.6-black.svg)](https://github.com/neurogenomics/MSTExplorer)
 [![](https://img.shields.io/github/languages/code-size/neurogenomics/MSTExplorer.svg)](https://github.com/neurogenomics/MSTExplorer)
 [![](https://img.shields.io/github/last-commit/neurogenomics/MSTExplorer.svg)](https://github.com/neurogenomics/MSTExplorer/commits/master)
 <br> [![R build