Alocating image sets to batches

Hi, I’d like to start by thanking you for developing CellProfiler, I still can’t believe such a useful piece of software is available for free. I’d also like to apologise for the length of this post in advance.

So, now my problem:
I am using high content microscopy to image slide samples of whole ex-vivo of embryonic mouse pancreas, each exposed to a different condition. My aim is to analyse the effects of each condition on amylase expression in the developing tissue. There are 2 tissues/conditions per slide and, using a template, I can image 4 slides per imaging run (see my layout image provided).
Unfortunately due to the irregular size, shape and placement of the tissues on each slide, I have to capture composite images of each tissue individually, with each tissue consisting of between 6-20 smaller images each.
Additionally the tissue is extremely dense so I have used a combination of E-cadherin and amylase immunofluorescence staining to identify the cell outlines in each tissue and save each outlined cell as an object. I then determine the intensity of amylase florescence within each object.

The problem comes when I want to process my images using cell profiler. If I process all the images I have collected together at once I end up with a huge database of tens of thousands of different objects with no efficient way of associating each image/object with each tissue. I have tried separating each set of images into batches by having the software recognise each position within my template as a different “well” and processing the images from each “well” separately. However because I have re-used the template for several sets of tissues, often two unrelated tissues will be processed together in the same “batch”.

What I want to know is, is there a way I can identify or label each set of images as belonging to a specific tissue, thus making it easier to pull them out of the database when I have finished processing.



E-cad identifier Amylase.cpproj (299.8 KB)Uploading…

Hello James!

Judging by your Metadata module, you relate metadata to your images based on File names. If the folders containing these images have information that relates to the conditions you investigated, you can extract this information to yield a unique “imaging run” tag, which can be used to distinguish the generic “A1-B4” nomenclature. For example:

Folder name “ImagingRun1” (and you know you imaged conditions control and 1-7)
Folder name “YouCalledItSomethingElseEntirely” (but you still know this contains conditions control and 8-14)

You can add “another extraction method” based on Folder name (don’t forget to add the tag to your database group columns!)


Ah brilliant thanks, I have added another metadata extraction and used a regular expression to identify the conditions (imaging run) from the path name.

You mentioned:

“don’t forget to add the tag to your database group columns!”

I assume this is an instruction to include my new “conditions” as a column in the database, is this correct? I’m not so hot on SQL stuff at the moment.

Exactly - I think, because I only know that it works, not why it works. I was assuming you want to use CP Analyst. If you choose to “create a CPA properties file” in the ExportToDatabase module, you can choose to group your data by setting “add group fields?” to yes. You can group your data on multiple levels, and it makes working with CPA easier.

If you are using other software I don’t think you need to do that.