Small UI update and bug fixes

* Changed the number of clusters from slider to numeric input
* Fixed some labeling bugs in a few of the figures
* More background changes to make code easier to read and maintain.

.gitignore

@@ -2,6 +2,7 @@
 .Rhistory
 .RData
 .Ruserdata
+*.Rproj
 
 .DS_Store
 

DESCRIPTION

@@ -1,6 +1,6 @@
 Package: JoesFlow
 Title: Joes Flow simplified analysis for single cell modality data
-Version: 0.1.1
+Version: 0.1.1-1
 Authors@R: c(person('Cooper', 'Devlin', email = '[email protected]', 
                     role = c('cre', 'aut')),
              person('Randy', 'Johnson', email = '[email protected]',

R/app_server.R

@@ -140,11 +140,10 @@ app_server <- function(input, output, session) {
 
   # Visualize::number of clusters
   output$cluster_setting<-renderUI({
-    sliderInput("kmean",
+    numericInput("kmean",
                 "Number of clusters:",
                 value = 5,
-                min = 1,
-                max = 20)
+                min = 2)
   })
 
 
@@ -271,8 +270,9 @@ app_server <- function(input, output, session) {
 
     gg <- pca_coords() %>%
       clusterJF(ids = data_mat()[,1],
-                meta = meta_mat()[,2],
-                colors = colors_samples)
+                meta = meta_mat()[,input$meta_val],
+                colors = colors_samples,
+                legend.name = input$meta_val)
 
     vals$pca_samps<-gg
 
@@ -287,7 +287,7 @@ app_server <- function(input, output, session) {
       clusterJF(ids = 1:nrow(data_mat()),
                 meta = kmeaner(),
                 colors = colors_clusters(),
-                legend.name = 'Kmeans')
+                legend.name = 'Cluster')
 
     vals$pca_kmeans<-gg
 
@@ -315,7 +315,8 @@ app_server <- function(input, output, session) {
                  ids = rownames(sb_pca()$groups_table),
                  meta = as.character(meta_mat()[,input$meta_val]),
                  colors1 = colors_samples,
-                 colors2 = colors_clusters())
+                 colors2 = colors_clusters(),
+                 legend.name = input$meta_val)
 
   })
   output$samp_p_pca <- renderPlot({
@@ -345,9 +346,11 @@ app_server <- function(input, output, session) {
   output$umap_plot = renderPlot({
 
     gg <- umap_coords() %>%
-      clusterJF(ids = data_mat()[,1],
-                meta = meta_mat()[,2],
-                colors = colors_samples)
+      clusterJF(axis_prefix = 'UMAP',
+                ids = data_mat()[,1],
+                meta = meta_mat()[,input$meta_val],
+                colors = colors_samples,
+                legend.name = input$meta_val)
 
     vals$umap_samps<-gg
 
@@ -358,10 +361,11 @@ app_server <- function(input, output, session) {
   output$umap_k_plot = renderPlot({
 
     gg <- umap_coords() %>%
-      clusterJF(ids = 1:nrow(data_mat()),
+      clusterJF(axis_prefix = 'UMAP',
+                ids = 1:nrow(data_mat()),
                 meta = kmeaner(),
                 colors = colors_clusters(),
-                legend.name = 'Kmeans')
+                legend.name = 'Cluster')
 
     vals$umap_kmeans<-gg
 
@@ -372,68 +376,41 @@ app_server <- function(input, output, session) {
 
   tsne_coords<-reactive({
     withProgress({
-      data_mat2=data_mat()[,-1]
-      ids=data_mat()[,1]
-      data_mat2=data.matrix(data_mat2)
-      mat = Rtsne::Rtsne(data_mat2, initial_dims=15, pca=TRUE, theta=1)$Y
+
+      mat <- data_mat()[,-1] %>%
+        Rtsne::Rtsne(initial_dims=15, pca=TRUE, theta=1) %>%
+        .[['Y']]
+
       colnames(mat)=c("tSNE_1", "tSNE_2")
+
       mat
 
     }, message="Calculating tSNE")
   })
 
   output$tsne_plot = renderPlot({
-    data_mat2=data_mat()[,-1]
-    ids=data_mat()[,1]
-    data_mat2=data.matrix(data_mat2)
-
-    tsne_df=tsne_coords()
-    plotter=data.frame(tSNE_1=tsne_df[,1], tSNE_2=tsne_df[,2], SampleID=ids)
-
-    if(length(input$meta_val)>0){
-      grouper=meta_mat()[,input$meta_val]
-
-      plotter$Group=as.character(plotter$SampleID)
-      samps=as.character(unique(plotter$SampleID))
-      for(jj in 1:length(samps)){
-        grouper=dplyr::filter(meta_mat(), .data$ID==samps[jj])
-        grouper=as.character(grouper[,input$meta_val][1])
-
-        plotter$Group[plotter$SampleID==samps[jj]]<-grouper
-      }
-    } else {
-      plotter$Group=plotter$SampleID
-    }
-
-    colors_samples2=colors_samples
-
-    gg=ggplot(sample(plotter), aes(.data$tSNE_1, .data$tSNE_2, color=.data$Group)) +
-      geom_point() + theme_bw() +
-      scale_color_manual(values=colors_samples2) +
-      theme(axis.text=element_text(color='black', size=14),
-            axis.title=element_text(color='black', size=16))
 
+    gg <- tsne_coords() %>%
+      clusterJF(axis_prefix = 'tSNE',
+                ids = data_mat()[,1],
+                meta = meta_mat()[,input$meta_val],
+                colors = colors_samples,
+                legend.name = input$meta_val)
+
     vals$tsne_samps<-gg
 
     print(gg)
 
   })
 
   output$tsne_k_plot = renderPlot({
-    data_mat2=data_mat()[,-1]
-    ids=data_mat()[,1]
-    data_mat2=data.matrix(data_mat2)
-
-    tsne_df=tsne_coords()
-    plotter=data.frame(tSNE_1=tsne_df[,1], tSNE_2=tsne_df[,2], SampleID=ids)
-
-    plotter$Kmeans=as.character(kmeaner())
-
-    gg=ggplot(sample(plotter), aes(.data$tSNE_1, .data$tSNE_2, color=.data$Kmeans)) +
-      geom_point() + theme_bw() +
-      scale_color_manual(values=colors_clusters()) +
-      theme(axis.text=element_text(color='black', size=14),
-            axis.title=element_text(color='black', size=16))
+
+    gg <- tsne_coords() %>%
+      clusterJF(axis_prefix = 'tSNE',
+                ids = 1:nrow(data_mat()),
+                meta = kmeaner(),
+                colors = colors_clusters(),
+                legend.name = 'Cluster')
 
     vals$tsne_kmeans<-gg
 

R/utils.R

@@ -7,6 +7,7 @@
 #' @param colors Vector of colors for each group
 #' @param xlab x-axis label
 #' @param ylab y-axis label
+#' @param axis_prefix x-and y-axis prefix (e.g. 'UMAP_')
 #' @param legend.name Character string for the legend name (default is 'Group')
 #' @param ... Other objects passed along to functions within clusterJF
 #' 
@@ -32,7 +33,7 @@ clusterJF.prcomp <- function(clustered_data, ids, meta, colors, legend.name = 'G
 
     # get group labels
     group_by(.data$SampleID) %>%
-    mutate(Group = meta[unique(.data$SampleID)]) %>%
+    mutate(Group = as.character(meta[unique(.data$SampleID)])) %>%
     ungroup()
 
   # proportion of variance
@@ -49,7 +50,7 @@ clusterJF.prcomp <- function(clustered_data, ids, meta, colors, legend.name = 'G
 #' @rdname clusterJF
 #' @method clusterJF matrix
 #' @export
-clusterJF.matrix <- function(clustered_data, ids, meta, colors, legend.name = 'Group', ...){
+clusterJF.matrix <- function(clustered_data, axis_prefix = 'axis', ids, meta, colors, legend.name = 'Group', ...){
 
   # format data for figure
   plotter <- tibble(X1       = clustered_data[,1],
@@ -58,11 +59,11 @@ clusterJF.matrix <- function(clustered_data, ids, meta, colors, legend.name = 'G
 
     # get group labels
     group_by(.data$SampleID) %>%
-    mutate(Group = meta[unique(.data$SampleID)]) %>%
+    mutate(Group = as.character(meta[unique(.data$SampleID)])) %>%
     ungroup()
 
   # render figure
-  clusterJF(plotter, colors, xlab = "UMAP_1", ylab = "UMAP_2", legend.name, ...)
+  clusterJF(plotter, colors, xlab = paste(axis_prefix, 1, sep = '_'), ylab = paste(axis_prefix, 2, sep = '_'), legend.name, ...)
 }
 
 # method for tibble object (should normally be called from clusterJF.prcomp or clusterJF.matrix)
@@ -90,14 +91,15 @@ clusterJF.tbl <- function(clustered_data, colors, xlab, ylab, legend.name, ...){
 #' @param meta Character vector containing metadata labels, corresponding to ids
 #' @param colors1 Vector of colors for samples
 #' @param colors2 Vector of colors for clusters
+#' @param legend.name Character string for the legend name (default is 'Group')
 #' 
 #' @return A ggplot object
 #' @export
 #' @import dplyr
 #' @import ggplot2
 #' @importFrom ggrepel geom_label_repel
 #' @importFrom rlang .data 
-sb_clusterJF <- function(clustered_data, ids, meta, colors1, colors2) {
+sb_clusterJF <- function(clustered_data, ids, meta, colors1, colors2, legend.name = 'Group') {
 
   # format data for figure
   plotter <- tibble(PC1      = clustered_data$x[,'PC1'], 
@@ -116,7 +118,7 @@ sb_clusterJF <- function(clustered_data, ids, meta, colors1, colors2) {
   pp1 <- ggplot(plotter, aes(.data$PC1, .data$PC2, color=.data$Group, label=.data$SampleID)) +
     geom_point() + theme_bw() +
     geom_label_repel(size = 6) +
-    scale_color_manual(values=colors1) +
+    scale_color_manual(values=colors1, name = legend.name) +
     guides(color = guide_legend(override.aes = list(label = 'O', size = 3))) +
     xlab(paste0("PC1 (Explained Variance ", round(PoV[1],4)*100, "%)")) +
     ylab(paste0("PC2 (Explained Variance ", round(PoV[2],4)*100, "%)")) +
@@ -135,7 +137,7 @@ sb_clusterJF <- function(clustered_data, ids, meta, colors1, colors2) {
   pp2 <- ggplot(plotter, aes(.data$PC1, .data$PC2, color=.data$Label, label=.data$Label)) +
     geom_point() + theme_bw() +
     geom_label_repel(size = 6) +
-    scale_color_manual(values=colors2) +
+    scale_color_manual(values=colors2, name = 'Cluster') +
     guides(color = guide_legend(override.aes = list(label = 'O', size = 3))) +
     theme(axis.text=element_text(color='black', size=14),
           axis.title=element_text(color='black', size=16))