Potential Cell Imaging/CellProfiler Week?

[cgreene]

We are considering adding this for 2018 or later. See https://github.com/CellProfiler/CellProfiler/issues/2664 for a very rough outline. Also we had some outstanding questions related to the CLI interface for CellProfiler.

[cgreene]

Potential structure:

Day 1 Prelab (before class):

• Install Cell Profiler and run some test (is there some test that you would use to just make sure the necessary components are installed?)

Day 1:

• Open CP
• Try some quick things with it (should they start with a base pipeline? do most people?).
• Figure out a series of steps that capture something desired in an image (maybe @AnneCarpenter has ideas for a good specific example)

Day 1 Homework:

• Save this as a pipeline

Day 2 Prelab:

• Run some standard pipeline on SageMathCloud using a command line version installed there (we'll need to loop in the SMC team to get it setup on their infrastructure).

Day 2:

• Use unix knowledge to run the pipeline over two different collections of images to generate summary statistics.

Day 2 Homework:

• ??? (TBD at this time)

Day 3 Prelab:

• Load ??? data into an R dataframe

Day 3 In Class:

• Analysis targeted to the specific question addressed by ??? data.

Day 3 Homework:

• Interpret results

[cgreene]

Defer consideration to 2019 version of the class.

[AnneCarpenter]

For Day 1 prelab, yes, you would install CellProfiler (executable would be ok, right? No reason to install from source?) and run the example pipeline: there is a welcome screen that pops up and suggests that you run an example and guides you through doing so. The skill level to do this is super minimal (if not from source code) so it should be fine for a prelab.

Day 1 lab you could have them pick any example pipeline + images from our examples page and run it: http://cellprofiler.org/examples/ (we might want to skim through to see if any should be excluded, for example, if the image set is too large or if it's some weird special case). Then, the open-ended task would be to measure something else interesting in the images (something beyond what the original pipeline was built for).

I think Day 2 & 3 could be based on David Julian's exercise, looking for differences in cell morphology between various cancer variants. I don't think his material is online but we can communicate to get that to you.