Trouble uplooading images - Oocytes counting

#1

Good Morning !

I’m having troubles uploading my pictures in CellProfiler. I don’t know if it is a format problem (I have tiff) or a size problem (around 900Ko) but when I drag my picture in the pipeline it freezes and never uploads . I attached an example of the picture I am working on.
My goal is to count the number of oocytes here and (that would be an absolute blast) classify them by size. I have a pipeline that have been working pretty well until now at least for the counting part (attached) but I haven’t attacked the classification problem yet and am now unable to load any picture anyway.
Any idea what I’ve been doing wrong ?

Thanks !

Marie

Oocyte_example.tiff (862.2 KB) Oocytes_Counting_CPF.cpproj (659.8 KB)

#2

Hi @MarieM27
The sample pipeline and image that you sent appears to work on my end with no problems. Are the images being stored on your local machine or are they on a network?

It might also be worth seeing if the same thing happens when you right click the image list in CellProfiler and use the “Browse for images” option to load in data instead of dragging.

#3

Hi Marie,

For me too there is no problem in loading your images. Check out the option that is been suggested by @DStirling.
Regarding you classification,
Once you segment Oocytes, you can compute its Area using “MeasureObjectSizeShape”. Then you can try using “ClassifyObjects” module where you can classify them based on your size measurements.

Hope this helps.

Regards,
Lakshmi
Fujifilm Wako Automation (Consultant)
www.wakoautomation.com
For Cellprofiler training or optimised pipeline write to,
lakshmi.balasubramanian.contractor@fujifilm.com

Join us at the High Content Analysis Symposium Oct 21-22, London
Society for Laboratory Automation and Screening

SLAS 2019 Advanced 3D Human Models and High-Content Analysis Symposium

Read more on our site.

Yokogawa CV8000 - The Ultimate in Confocal HCS
https://www.wakoautomation.com/products/yokogawa-high-content-imaging

#4

Hi !
Thanks for the answers. I tried the solution proposed using “Browse for image” instead of dragging and it’s not working better. I got it to work once and then it stopped working again :(. All the pictures are stored on my laptop though. As I’m still stuck with no pictures uploading I didn’t get to try your solution @lakshmi.

Thanks again for your help,

Marie

#5

Hi @MarieM27,

Could you confirm which platform and version of CellProfiler you’re using, and whether the image you uploaded above is definitely the one causing the crash (try just loading that alone)? Also, if you’re on Windows does anything appear on the black console window when you try to open the file?

Thanks

#6

Hi !
So the image I’m using is not responsible for the crash, I tried with a couple of them and none seem to work. I’m using a MacBook Air and regarding the CellProfiler version I’m using 3.1.8.

Thanks @DStirling for taking the time to help me troubleshoot this,

Marie

#7

Hi @MarieM27

That does sound like an issue with the program itself rather than the file format. Could you try installing the newest version (3.1.9) to see if that fixes it?

#8

Hi again,

So I uploaded the last version of the software and it solved the problem for one of the pictures but not for all of them. I attached a couple of the non working ones. MIS 10x-1 (glissées).tiff (1.6 MB) MIS back 10x-1 (glissées).tiff (1.1 MB)

Thanks

Marie

#9

Hi @MarieM27,

Thanks for that, I’m able to replicate your issue. It appears that you need to remove the accented characters from the file names (the é) to get CellProfiler to acknowledge them. I’ll look into whether this is a bug that we can fix for a future version.

#10

@MarieM27 - Having investigated, it looks like such characters shouldn’t cause a problem in future versions of CellProfiler, but for now you’ll have to remove them from both the image file names and the directories they’re stored in. Hope that helps!

#11

Thanks ! The uploading problem is fixed now :).
Can I push my luck here and ask for adivces regarding the pipeline itself. It’s working, more or less, but because I can’t use DAPI to count the nucleus and the oocytes intensity is not uniform I struggle to count them correctly. I tried to play on the size of the objects detected but it’s not helping.
Any suggestions ?

Thanks again,

Marie

#12

I attached the pipeline here.

Oocytes_Counting_CPF2.cpproj (647.5 KB)

#13

Hi @MarieM27,

I tried your pipeline. I have just added an median filter before your segmentation step. This is the result,
oocytes_output.pdf (295.6 KB) .
This may not be perfect but definitely one way to improve the segmentation.
Please find your pipeline,
Oocytes_Counting_CPF2.cpproj (651.3 KB)

Regards,
Lakshmi
Fujifilm Wako Automation (Consultant)
www.wakoautomation.com
For CellProfiler training or optimised pipeline write to,
lakshmi.balasubramanian.contractor@fujifilm.com

Read more on our site.
Yokogawa CV8000 - The Ultimate in Confocal HCS
https://www.wakoautomation.com/products/yokogawa-high-content-imaging

#14

Hi @lakshmi,

Thanks for your help, it’s already way better for several pictures. I still have problems when the intensity varies between the different objects I’m trying to detect. Do you think it’s possible to improve the pipeline to solve this issue ?
I attached an example.

Thanks again !

Marie

CON 9.tif (2.2 MB)

19
Capture d’écran, le 2019-11-05 à 16.32.19.png2880×1760 451 KB

#15

Hi @MarieM27,

Not sure. In the regions where the segmentation is not fine, the signal is high and it is as same as background. Just I tried changing the thresholding to three class which improved slight. You could try. Please find the screenshot here,

oocytes_improvement
oocytes_improvement1920×1080 251 KB

Regards,
Lakshmi
Fujifilm Wako Automation (Consultant)
www.wakoautomation.com
For CellProfiler training or optimised pipeline write to,
lakshmi.balasubramanian.contractor@fujifilm.com

Read more on our site.
Yokogawa CV8000 - The Ultimate in Confocal HCS
https://www.wakoautomation.com/products/yokogawa-high-content-imaging

Join us at, https://www.slas2020.org