Update auto_crop.R

Hi Lucy. I put a couple of comments and queries. Let me know if this works and makes sense. Best, Wayne

source/auto_crop.R

+#' A cropping function
+#'
+#' This function identifies meiotic nuclei and crops them.
+#' @keywords synapsis
+#' @export
+#' @examples
+#' To use the "crop" function on a folder of images, apply the following three lines of code
+#' path = "/folder_of_your_images"
+#' files <- list.files(path)
+#' crop(files, "/folder_of_your_images")
+#' 
+#' # set the path below to where your jpegs are located
+#' # to save the cropped images, make a folder called "crops" which in "/folder_of_your_images" in other words "/folder_of_your_images/crops"
 crop <- function(file_list, img_path, crop_method = "regular")
 {
   # input :
@@ -11,8 +24,8 @@ crop <- function(file_list, img_path, crop_method = "regular")
   cell_count <- 0
   image_count <-0
   pair <- 0
-
-  ## for each image that is *-dna.jpeg,
+# I think that what is being referred to as ".dna" below should be referred to as "sycp3", or perhaps "axis". Happy to discuss more about this important meiosis-nerd distinction. 
+  ## Pairing multiple channels for each image e.g. the 594 axis marker and the 488 recombination marker 
   for (file in file_list){
     setwd(img_path)
     if(grepl("*dna.jpeg$", file)){
@@ -31,7 +44,7 @@ crop <- function(file_list, img_path, crop_method = "regular")
     }
     if(pair ==1){
 
-      #### function: blur the image
+      #### function: blur the image 
       ## call it on img_orig, optional offset
       blob_th <- get_blobs(img_orig)
 
@@ -39,8 +52,7 @@ crop <- function(file_list, img_path, crop_method = "regular")
       blob_label <- channel(blob_label, "gray")
       candidate <- bwlabel(blob_label)
 
-      ## function: remove things that aren't cells
-
+      ## function: remove things that aren't cells (#I am confused about something called "keep cells" being assigned to a variable called "removed".)
       removed <- keep_cells(candidate)
 
       ### crop over each cell
@@ -81,6 +93,7 @@ print("viable cells")
 ### add all the functions used here
 
 #################################### new function ####################################
+# this function blurs nuclei that have been stained with antibodies against axis proteins
 get_blobs <- function(img_orig, crop_method = "regular"){
 
   # input:
@@ -110,6 +123,7 @@ get_blobs <- function(img_orig, crop_method = "regular"){
 }
 
 #################################### new function ####################################
+# this function does "ABC"
 keep_cells <- function(candidate){
 
   # input:
@@ -120,12 +134,14 @@ keep_cells <- function(candidate){
 
   # delete everything that's too small
   colorimg<- colorLabels(candidate, normalize = TRUE)
-  x <- computeFeatures.shape(candidate)
+  x <- computeFeatures.shape(candidate) # uses function from ebimage. give output kind of related to how circular the object is, size, perimeter, radius average etc
   x <- data.frame(x)
-  OOI <- width(x)
+  OOI <- width(x)  #Object Of Interest 
   counter <- 0
   removed <- candidate
 
+# loops over each Object of Interest (OOI)
+
   while(counter<OOI){
     counter <- counter+1
     pixel_area = x$s.area[counter]
@@ -145,7 +161,7 @@ keep_cells <- function(candidate){
   return(removed)
 }
 
-
+# function to help not count the same object twice
 crop_single_object <- function(removed, OOI_final,counter_final,img_orig,img_orig_foci,file,cell_count){
   tmp_img <- removed
   ## have a single object
@@ -162,6 +178,7 @@ crop_single_object <- function(removed, OOI_final,counter_final,img_orig,img_ori
 
   }
 
+# put nucleus mask on original image of the axis-marker or foci channel. matrix stuff which crops it to a square
   ## function: remove noise
   noise_gone <- bwlabel(tmp_img)*as.matrix(img_orig)
   noise_gone_foci <- bwlabel(tmp_img)*as.matrix(img_orig_foci)