Use newly computed metrics

DESCRIPTION

@@ -15,7 +15,7 @@ Description: This package contains precomputed data including comparisons in
 License: AGPL (>= 3)
 Encoding: UTF-8
 Roxygen: list(markdown = TRUE)
-RoxygenNote: 7.2.0
+RoxygenNote: 7.2.1
 Depends:
     R (>= 2.10)
 LazyData: true

R/genes_table.R

@@ -71,14 +71,17 @@ genes_table_server <- function(id, data) {
             "target=\"_blank\">{hgnc_name}</a>"
           ),
           "Rank" = rank,
-          "Percentile" = percentile,
+          "%" = percentile,
           "Score" = score,
           "Median" = median_expression,
+          "IQR" = iqr_expression,
+          "QCV" = qcv_expression,
           "Mean" = mean_expression,
-          "Standard deviation" = sd_expression,
+          "SD" = sd_expression,
-          "Expressed" = above_zero,
+          "CV" = cv_expression,
-          "Above median" = above_median,
+          "> 0" = above_zero,
-          "Above 95%" = above_95
+          "> median" = above_median,
+          "> 95%" = above_95
         )],
         options = list(
           dom = "frtip",
@@ -90,18 +93,21 @@ genes_table_server <- function(id, data) {
       ) |>
         DT::formatPercentage(
           c(
-            "Percentile",
+            "%",
-            "Score",
+            "> 0",
-            "Expressed",
+            "> median",
-            "Above median",
+            "> 95%"
-            "Above 95%"
           ),
           digits = 2,
         ) |>
         DT::formatRound(c(
+          "Score",
           "Median",
+          "IQR",
+          "QCV",
           "Mean",
-          "Standard deviation"
+          "SD",
+          "CV"
         ))
     })
   })

R/ranking.R

@@ -3,34 +3,26 @@
 #' 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,
+                       level_metric = "median_expression_normalized",
-                       sd_expression_weight = -0.25) {
+                       level_weight = 0.25,
+                       variation_metric = "qcv_expression_normalized",
+                       variation_weight = -0.25) {
   total_weight <- abs(cross_sample_weight) +
-    abs(mean_expression_weight) +
+    abs(level_weight) +
-    abs(sd_expression_weight)
+    abs(variation_weight)
 
   data <- copy(ubigen::genes)
 
   data[, score :=
     (cross_sample_weight * get(cross_sample_metric) +
-      mean_expression_weight * mean_expression_normalized +
+      level_weight * get(level_metric) +
-      sd_expression_weight * sd_expression_normalized) /
+      variation_weight * get(variation_metric)) /
       total_weight]
 
   # Normalize scores to be between 0.0 and 1.0.

R/server.R

@@ -5,8 +5,10 @@ server <- function(input, output, session) {
     rank_genes(
       cross_sample_metric = input$cross_sample_metric,
       cross_sample_weight = input$cross_sample_weight,
-      mean_expression_weight = input$mean_expression,
+      level_metric = input$level_metric,
-      sd_expression_weight = input$sd_expression
+      level_weight = input$level_weight,
+      variation_metric = input$variation_metric,
+      variation_weight = input$variation_weight
     )
   })
 

R/ui.R

@@ -43,26 +43,45 @@ ui <- function() {
               step = 0.01,
               value = 0.5
             ),
-            sliderInput(
+            selectInput(
-              "mean_expression",
+              "level_metric",
               verticalLayout(
-                strong("Mean Expression"),
+                strong("Expression level"),
-                "Mean expression of the gene across all samples."
+                "Typical expression level of the gene across all samples."
               ),
+              list(
+                "Median expression" = "median_expression_normalized",
+                "Mean expression" = "mean_expression_normalized"
+              )
+            ),
+            sliderInput(
+              "level_weight",
+              label = NULL,
               min = -1.0,
               max = 1.0,
               step = 0.01,
               value = 0.25
             ),
-            sliderInput(
+            selectInput(
-              "sd_expression",
+              "variation_metric",
               verticalLayout(
-                strong("Standard deviation"),
+                strong("Expression variation"),
                 paste0(
-                  "Standard deviation of the gene's expression across all ",
+                  "Measure of the variation of the gene's expression between ",
                   "samples."
                 )
               ),
+              list(
+                "Quantile based coefficient of variation" =
+                  "qcv_expression_normalized",
+                "Interquartile range" = "iqr_expression_normalized",
+                "Coefficient of variation" = "cv_expression_normalized",
+                "Standard deviation" = "sd_expression_normalized"
+              )
+            ),
+            sliderInput(
+              "variation_weight",
+              label = NULL,
               min = -1.0,
               max = 1.0,
               step = 0.01,

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 14 columns.
+An object of class \code{data.table} (inherits from \code{data.frame}) with 56156 rows and 20 columns.
 }
 \usage{
 genes

man/rank_genes.Rd

@@ -7,24 +7,12 @@
 rank_genes(
   cross_sample_metric = "above_95",
   cross_sample_weight = 0.5,
-  mean_expression_weight = 0.25,
+  level_metric = "median_expression_normalized",
-  sd_expression_weight = -0.25
+  level_weight = 0.25,
+  variation_metric = "qcv_expression_normalized",
+  variation_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.

scripts/gsea.R

@@ -42,8 +42,8 @@ fig <- plotly::plot_ly(data) |>
     color = ~source
   ) |>
   plotly::add_lines(
-    x = ~bucket_smoothed,
+    x = bucket_smoothed,
-    y = ~total_smoothed,
+    y = total_smoothed,
     name = "All (smoothed)"
   ) |>
   plotly::layout(