Will it be possible in the future to make boxplots and histograms of negative vs positive objects from all images in an experiment? I have been making the boxplot figures of my positive and negative scored class objects in R but it is another analysis step and if the functionality is in CPA already I might as well use it. Also, I would like to save a couple images of scored positive and negative control tissues but the class label overlay does not show up if I try to save the image in CPA. I have been taking screenshots with print screen up till now. Is there a way to save the class labeled image in CPA without using print screen?
This functionality will be added in the future, but at the moment we don’t really have the time (funding) to add many more features to CPA.
There’s an imperfect workaround however that will allow you to filter out object of a particular class for plotting (but not to plot each class at the same time against each other):
- set class_table = my_class_table in your properties file, this will cause classifier to write out per-object classifications to “my_class_table”
- after you’ve classified your cells and created the class table, open the histogram and plot my_class_table.class_number (you’ll probably have to select OTHER TABLE and link the class table to your per_object table via the ImageNumber and ObjectNumber columns)
- create a new gate (via the gate dropdown) named “positive” and draw a selection around the “positive” objects (or objects of any class of interest)
- open another histogram (the other tools will work too) plot and plot some measurement from the per_object table, and set the filter to “positive”
…now you can see the per-cell distribution of that measurement for your “positive” objects.
This is another feature that we’ve had requests for but just haven’t had the time to implement for the reasons listed above. Unfortunately, screenshots will have to suffice until we (or someone else) finds the time to add this feature.
Thanks for your questions,