Update gene data based on the suggested ranking

This also exports the ranking function itself for external use.

NAMESPACE

@@ -1,5 +1,6 @@
 # Generated by roxygen2: do not edit by hand
 
+export(rank_genes)
 export(run_app)
 import(data.table)
 import(shiny)

R/ranking.R

@@ -0,0 +1,48 @@
+#' Rank genes based on how ubiquitous they are.
+#'
+#' This function will compute a weighted average across multiple metrics that
+#' define how ubiquitous a gene is based on its expression across samples.
+#'
+#' @param cross_sample_metric Metric to use for calculating the number of
+#'   samples a gene is expressed in. One of `above_95`, `above_median` or
+#'   `above_zero`.
+#' @param cross_sample_weight Weighting of the cross sample metric within the
+#'   final score.
+#' @param mean_expression_weight Weighting of the gene's mean expression within
+#'   the final score.
+#' @param sd_expression_weight Weighting of the standard deviation of the
+#'   gene's expression within the final score.
+#'
+#' @return A `data.table` with gene data as well as the scores, ranks and
+#'   percentiles for each gene.
+#'
+#' @export
+rank_genes <- function(cross_sample_metric = "above_95",
+                       cross_sample_weight = 0.5,
+                       mean_expression_weight = 0.25,
+                       sd_expression_weight = -0.25) {
+  total_weight <- cross_sample_weight +
+    mean_expression_weight +
+    sd_expression_weight
+
+  data <- copy(ubigen::genes)
+
+  data[, score :=
+    (cross_sample_weight * get(cross_sample_metric) +
+      mean_expression_weight * mean_expression_normalized +
+      sd_expression_weight * sd_expression_normalized) /
+      total_weight]
+
+  # Normalize scores to be between 0.0 and 1.0.
+  data[, score := (score - min(score, na.rm = TRUE)) /
+    (max(score, na.rm = TRUE) - min(score, na.rm = TRUE))]
+
+  # These are genes that are not expressed at all.
+  data[is.na(score), score := 0.0]
+
+  setorder(data, -score)
+  data[, rank := .I]
+  data[, percentile := 1 - rank / max(rank)]
+
+  data
+}

R/server.R

@@ -2,30 +2,12 @@
 #' @noRd
 server <- function(input, output, session) {
   ranked_data <- reactive({
-    total_weight <- abs(input$cross_sample_weight) +
+    rank_genes(
-      abs(input$mean_expression) +
+      cross_sample_metric = input$cross_sample_metric,
-      abs(input$sd_expression)
+      cross_sample_weight = input$cross_sample_weight,
-
+      mean_expression_weight = input$mean_expression,
-    data <- data.table::copy(ubigen::genes)
+      sd_expression_weight = input$sd_expression
-
+    )
-    data[, score :=
-      (input$cross_sample_weight * get(input$cross_sample_metric) +
-        input$mean_expression * mean_expression_normalized +
-        input$sd_expression * sd_expression_normalized) /
-        total_weight]
-
-    # Normalize scores to be between 0.0 and 1.0.
-    data[, score := (score - min(score, na.rm = TRUE)) /
-      (max(score, na.rm = TRUE) - min(score, na.rm = TRUE))]
-
-    # These are genes that are not expressed at all.
-    data[is.na(score), score := 0.0]
-
-    data.table::setorder(data, -score)
-    data[, rank := .I]
-    data[, percentile := 1 - rank / max(rank)]
-
-    data
   })
 
   custom_genes <- gene_selector_server("custom_genes") |> debounce(500)

man/genes.Rd

@@ -5,7 +5,7 @@
 \alias{genes}
 \title{A \code{data.table} containig data on genes and their expression behavior.}
 \format{
-An object of class \code{data.table} (inherits from \code{data.frame}) with 56156 rows and 13 columns.
+An object of class \code{data.table} (inherits from \code{data.frame}) with 56156 rows and 14 columns.
 }
 \usage{
 genes

man/rank_genes.Rd

@@ -0,0 +1,35 @@
+% Generated by roxygen2: do not edit by hand
+% Please edit documentation in R/ranking.R
+\name{rank_genes}
+\alias{rank_genes}
+\title{Rank genes based on how ubiquitous they are.}
+\usage{
+rank_genes(
+  cross_sample_metric = "above_95",
+  cross_sample_weight = 0.5,
+  mean_expression_weight = 0.25,
+  sd_expression_weight = -0.25
+)
+}
+\arguments{
+\item{cross_sample_metric}{Metric to use for calculating the number of
+samples a gene is expressed in. One of \code{above_95}, \code{above_median} or
+\code{above_zero}.}
+
+\item{cross_sample_weight}{Weighting of the cross sample metric within the
+final score.}
+
+\item{mean_expression_weight}{Weighting of the gene's mean expression within
+the final score.}
+
+\item{sd_expression_weight}{Weighting of the standard deviation of the
+gene's expression within the final score.}
+}
+\value{
+A \code{data.table} with gene data as well as the scores, ranks and
+percentiles for each gene.
+}
+\description{
+This function will compute a weighted average across multiple metrics that
+define how ubiquitous a gene is based on its expression across samples.
+}