added create_nodedf function, small changes to browser

DESCRIPTION

@@ -11,7 +11,7 @@ Encoding: UTF-8
 LazyData: true
 biocViews:
 Imports:
-RoxygenNote: 7.1.1
+RoxygenNote: 7.1.2
 Suggests: 
     clusterProfiler (>= 3.10.0),
     mgcv (>= 1.8.24),

NAMESPACE

@@ -5,6 +5,7 @@ export(add_pw_protection)
 export(arrayqc)
 export(calc_logfc)
 export(calc_logfc_public)
+export(create_nodedf)
 export(create_nodelist)
 export(create_nodelist_voom)
 export(create_shiny_mapplotlist)

R/create_nodelist.R

 #' Create logFC Nodelist
 #'
+#' A simple wrapper function to applying create_nodedf() to all nodes in a nodeframe.
+#'
 #' @param elist An limma EList object with normalized logFC values
 #' @param nodeframe a nodeframe describing the associations of genes in the toxicogenomic universe
 #' @param keep_recovery (Boolean) Defaults to False, intended for tcta fit, set to True if recoveries should be analysed
@@ -18,100 +20,67 @@ create_nodelist <- function(elist, nodeframe, keep_recovery = F) {
   } # for smoothed datasets
 
   # create list with logFC and metadata for each node -----------------------
-
-  if (!keep_recovery) {
-    nodelist <-
+  nodelist <-
       lapply(
-        X = seq(1, max(nodeframe_elist$node)),
-        FUN = function(nodeID) {
-          if (sum(nodeframe_elist$node == nodeID) > 0) {
-            logFC <-
-              c(t(elist$E[elist$genes$ProbeName %in% nodeframe_elist$ProbeID[nodeframe_elist$node == nodeID], elist$targets$type != "recovery"]))
-            concentration <-
-              rep(
-                elist$targets$concentration[elist$targets$type != "recovery"],
-                times = sum(nodeframe_elist$node == nodeID)
-              )
-            concentration_level <-
-              rep(
-                elist$targets$concentration_level[elist$targets$type != "recovery"],
-                times = sum(nodeframe_elist$node == nodeID)
-              )
-            time_hpe_factor <-
-              ordered(rep(
-                elist$targets$time_hpe[elist$targets$type != "recovery"],
-                times = sum(nodeframe_elist$node == nodeID)
-              ))
-            time_hpe <-
-              rep(elist$targets$time_hpe[elist$targets$type != "recovery"],
-                  times = sum(nodeframe_elist$node == nodeID))
-            probe_id <-
-              rep(as.character(nodeframe_elist$ProbeID)[nodeframe_elist$node == nodeID], each = nrow(t(elist$E[elist$genes$ProbeName %in% nodeframe_elist$ProbeID[nodeframe_elist$node == nodeID], elist$targets$type != "recovery"])))
-            ensembl_gene_id <-
-              rep(as.character(nodeframe_elist$ensembl)[nodeframe_elist$node == nodeID], each = nrow(t(elist$E[elist$genes$ProbeName %in% nodeframe_elist$ProbeID[nodeframe_elist$node == nodeID], elist$targets$type != "recovery"])))
-            substance <- elist$targets$substance[1]
-            type <-
-              rep(
-                elist$targets$type[elist$targets$type != "recovery"],
-                times = sum(nodeframe_elist$node == nodeID)
-              )
-            probeframe <-
-              data.frame(
-                logFC = logFC,
-                concentration = concentration,
-                concentration_level = concentration_level,
-                time_hpe_factor = time_hpe_factor,
-                time_hpe = time_hpe,
-                probe_id = probe_id,
-                ensembl_gene_id = ensembl_gene_id,
-                nodeID = nodeID,
-                substance = substance,
-                type = type
-              )
-            return(probeframe)
-          } else {
-            return(NA)
-          }
-        }
+          X = seq(1, max(nodeframe_elist$node)),
+          FUN = create_nodedf,
+          elist = elist,
+          nodeframe = nodeframe,
+          keep_recovery = keep_recovery
       )
-  } else {
-    nodelist <-
-      lapply(
-        X = seq(1, max(nodeframe_elist$node)),
-        FUN = function(nodeID) {
-          if (sum(nodeframe_elist$node == nodeID) > 0) {
-            logFC <-
-              c(t(elist$E[elist$genes$ProbeName %in% nodeframe_elist$ProbeID[nodeframe_elist$node == nodeID],]))
-            concentration <-
-              rep(
-                elist$targets$concentration,
-                times = sum(nodeframe_elist$node == nodeID)
-              )
-            concentration_level <-
-              rep(
-                elist$targets$concentration_level,
-                times = sum(nodeframe_elist$node == nodeID)
-              )
-            time_hpe_factor <-
-              ordered(rep(
+  nodelist
+}
+
+#' Create a nodedf
+#'
+#' Creates a data frame with logFC and metadata for a single node
+#'
+#' @param nodeID id of the node
+#' @param elist limma elist
+#' @param nodeframe a nodeframe describing the associations of genes in the toxicogenomic universe
+#' @param keep_recovery (Boolean) Defaults to False, intended for tcta fit, set to True if recoveries should be analysed
+#'
+#' @return a data frame or NA if the node does noch contain any genes
+#' @export
+#'
+create_nodedf = function(nodeID, elist, nodeframe, keep_recovery = F) {
+    nodeframe$ensembl <- as.character(nodeframe$ensembl)
+    nodeframe$ProbeID <- as.character(nodeframe$ProbeID)
+    elist$genes$ProbeName <- as.character(elist$genes$ProbeName)
+
+    nodeframe_elist <- nodeframe[!is.na(nodeframe$ProbeID), ]
+
+    if (is.null(elist$targets$type)) {
+        elist$targets$type <- "treatment"
+    } # for smoothed datasets
+
+    if (sum(nodeframe_elist$node == nodeID) > 0) {
+        logFC <-
+            c(t(elist$E[elist$genes$ProbeName %in% nodeframe_elist$ProbeID[nodeframe_elist$node == nodeID], ]))
+        concentration <-
+            rep(elist$targets$concentration,
+                times = sum(nodeframe_elist$node == nodeID))
+        concentration_level <-
+            rep(elist$targets$concentration_level,
+                times = sum(nodeframe_elist$node == nodeID))
+        time_hpe_factor <-
+            ordered(rep(
                 elist$targets$time_hpe,
                 times = sum(nodeframe_elist$node == nodeID)
-              ))
-            time_hpe <-
-              rep(elist$targets$time_hpe,
-                  times = sum(nodeframe_elist$node == nodeID))
-            probe_id <-
-              rep(as.character(nodeframe_elist$ProbeID)[nodeframe_elist$node == nodeID], each = nrow(t(elist$E[elist$genes$ProbeName %in% nodeframe_elist$ProbeID[nodeframe_elist$node == nodeID],])))
-            ensembl_gene_id <-
-              rep(as.character(nodeframe_elist$ensembl)[nodeframe_elist$node == nodeID], each = nrow(t(elist$E[elist$genes$ProbeName %in% nodeframe_elist$ProbeID[nodeframe_elist$node == nodeID],])))
-            substance <- elist$targets$substance[1]
-            type <-
-              rep(
-                elist$targets$type,
-                times = sum(nodeframe_elist$node == nodeID)
-              )
-            probeframe <-
-              data.frame(
+            ))
+        time_hpe <-
+            rep(elist$targets$time_hpe,
+                times = sum(nodeframe_elist$node == nodeID))
+        probe_id <-
+            rep(elist$genes$ProbeName[elist$genes$ProbeName %in% nodeframe_elist$ProbeID[nodeframe_elist$node == nodeID]], each = nrow(t(elist$E[elist$genes$ProbeName %in% nodeframe_elist$ProbeID[nodeframe_elist$node == nodeID], ])))
+        ensembl_gene_id <-
+            rep(elist$genes$ensembl_gene_id[elist$genes$ProbeName %in% nodeframe_elist$ProbeID[nodeframe_elist$node == nodeID]], each = nrow(t(elist$E[elist$genes$ProbeName %in% nodeframe_elist$ProbeID[nodeframe_elist$node == nodeID], ])))
+        substance <- elist$targets$substance[1]
+        type <-
+            rep(elist$targets$type,
+                times = sum(nodeframe_elist$node == nodeID))
+        probeframe <-
+            data.frame(
                 logFC = logFC,
                 concentration = concentration,
                 concentration_level = concentration_level,
@@ -122,15 +91,14 @@ create_nodelist <- function(elist, nodeframe, keep_recovery = F) {
                 nodeID = nodeID,
                 substance = substance,
                 type = type
-              )
-            return(probeframe)
-          } else {
-            return(NA)
-          }
+            )
+        if (!keep_recovery) {
+            probeframe = probeframe[probeframe$type != 'recovery',]
         }
-      )
-  }
-  nodelist
+        return(probeframe)
+    } else {
+        return(NA)
+    }
 }
 
 

R/data_for_shiny.R

@@ -60,7 +60,8 @@ create_shiny_mapplotlist <-
             'group',
             'model',
             'url',
-            'condition'
+            'condition',
+            'unit'
         )]
     }) %>% do.call('rbind', .) %>%
         unique()

R/import_data.R

@@ -259,7 +259,7 @@ import_seq_data = function(counts,
     keep = rowSums(counts) >= ncol(counts) - 1
     dge = dge[keep,]
 
-    voom = voomWithQualityWeights(
+    voom = limma::voomWithQualityWeights(
         counts = dge,
         normalize.method = 'cyclicloess'
     )

R/plottimepoint.R

@@ -5,7 +5,7 @@
 #' @return a ggplot
 #' @export
 #'
-plot_noderesponse = function(plot_data) {
+plot_noderesponse = function(plot_data, conc_unit = 'µM') {
     model = plot_data$model
 
     nodeplot_theme = theme_bw() +
@@ -59,7 +59,7 @@ plot_noderesponse = function(plot_data) {
         ) +
         scale_colour_discrete(name = "Gene ID") +
         ylab("logFC") +
-        xlab("\n\nexposure concentration [µM]") +
+        xlab(sprintf('\nexposure concentration [%s]', conc_unit)) +
         {
             if (model == 'lm') {
                 # the limits are extended by 20% of the max concentration so the

R/tfp_browser.R

@@ -30,6 +30,24 @@ tfp_browser = function(filename, version = '4.0.0') {
         return(sort(round(breaks, digits = 1)))
     }
 
+    convert_unit = function(str) {
+        if (str_detect(str, '^log')) {
+            log = T
+        } else {
+            log = F
+        }
+        str = str_remove(str, '^log\\d+_')
+        if (str_detect(str, '_L$')) {
+            str = str_remove(str, '_L$')
+            str = paste0(str, '/L')
+        }
+        if (log) {
+            str = paste0('log(', str, ')')
+        }
+        str = str_replace(str, 'u', 'µ')
+        return(str)
+    }
+
     # function for plotting text in the space where the SOM would be
     generic_plot = function(text, font_size = 5) {
         ggplot2::ggplot(data = data.frame(x = 1,

R/tfp_server.R

@@ -459,8 +459,12 @@ tfp_server = function(input, output, session) {
 
             output$fingerprintheader = renderText({
                 sprintf(
-                    'Transcriptomic response for %s at %.2f μmol/L and %s hpe',
+                    'Transcriptomic response for %s at %.2f %s and %s hpe',
                     input$selectedsubstance1,
+                    targets()$unit[
+                        targets()$group == input$selectedgroup &
+                            targets()$substance == input$selectedsubstance1 &
+                            targets()$concentration_level == input$concselect1][1] %>% convert_unit(),
                     targets()$concentration[
                         targets()$group == input$selectedgroup &
                         targets()$substance == input$selectedsubstance1 &
@@ -482,6 +486,10 @@ tfp_server = function(input, output, session) {
                     selected_group = input$selectedgroup
                     selected_condition = input$selectedcondition1
 
+                    unit = targets()$unit[targets()$group == selected_group &
+                                              targets()$substance == selected_substance &
+                                              targets()$concentration_level == input$concselect1][1] %>% convert_unit()
+
                     D_measured = tbl(db, 'D_measured_all') %>%
                         filter(group == selected_group &
                                    substance == selected_substance &
@@ -510,7 +518,7 @@ tfp_server = function(input, output, session) {
                     )
 
                     if (dim(D_measured)[1] > 0) {
-                        plot_noderesponse(plot_data = plot_data)
+                        plot_noderesponse(plot_data = plot_data, conc_unit = unit)
                     } else {
                         # selected node has no data
                         generic_plot('No measured data for selected node.\n Please select different node.')

man/create_nodedf.Rd

+% Generated by roxygen2: do not edit by hand
+% Please edit documentation in R/create_nodelist.R
+\name{create_nodedf}
+\alias{create_nodedf}
+\title{Create a nodedf}
+\usage{
+create_nodedf(nodeID, elist, nodeframe, keep_recovery = F)
+}
+\arguments{
+\item{nodeID}{id of the node}
+
+\item{elist}{limma elist}
+
+\item{nodeframe}{a nodeframe describing the associations of genes in the toxicogenomic universe}
+
+\item{keep_recovery}{(Boolean) Defaults to False, intended for tcta fit, set to True if recoveries should be analysed}
+}
+\value{
+a data frame or NA if the node does noch contain any genes
+}
+\description{
+Creates a data frame with logFC and metadata for a single node
+}

man/create_nodelist.Rd

@@ -17,5 +17,5 @@ create_nodelist(elist, nodeframe, keep_recovery = F)
 A List with logFC values for each node of the toxicogenomic universe
 }
 \description{
-Create logFC Nodelist
+A simple wrapper function to applying create_nodedf() to all nodes in a nodeframe.
 }

man/plot_noderesponse.Rd

@@ -4,7 +4,7 @@
 \alias{plot_noderesponse}
 \title{Plot noderesponse over all time points}
 \usage{
-plot_noderesponse(plot_data)
+plot_noderesponse(plot_data, conc_unit = "µM")
 }
 \arguments{
 \item{plot_data}{output from get_nodeplot_data()}