Describe expected appearance of plot_stability.

.Rbuildignore

@@ -0,0 +1,14 @@
+^.*\.Rproj$
+^\.Rproj\.user$
+^Makefile$
+^development$
+^doc$
+^test$
+^test_output$
+^paper
+^_pkgdown.yml$
+^index.md$
+^codemeta.json$
+^docs$
+^varistran.sublime-
+

R/plot.R

@@ -70,6 +70,8 @@
 #' Genes are partitioned evenly into "bins" bins by average expression level.
 #' Mean residual standard deviation is plotted against mean count.
 #'
+#' If the variance stabilizing transformation has been successful, this plot should be close to a horizontal line. However it is normal for the standard deviation to drop off for counts below 5.
+#'
 #' @param y Transformed counts matrix.
 #' @param x Optional, original counts matrix.
 #' @param design Matrix specifying a linear model with which to calculate

R/shiny_plots.R

@@ -47,7 +47,8 @@ shiny_stability <- function(y, x=NULL, design=NULL, bins=20, prefix="") {
             "It shows the amount of noise in the transformed data ",
             "across different expression levels. ",
             "If the transformation has been successful, ",
-            "this should be close to a flat line."
+            "this should be close to a flat line.",
+            "However the standard deviation will usually drop off for mean counts below 5."
             ),
         parenthetically("This plot is produced by varistran::plot_stability.")
     )