Only loading and processing half of slices from an stk file?

Hello,

Disclaimer: I am not sure whether this is an error on my end or something to do with the software, so I’m putting this here in the Help section.

I’ve been using CP and Analyst for some time now while working with conventional, 3 channel microscopy images. Recently, I moved to a new project where we are using an MDS IXM to automate the process of acquiring z-stacks of cells printed in ~60 nL alginate blobs. The native files output by the device are stk, and so I’ve been trying to use Load Images with the stk option. Everything seems to work swimmingly except for one hang-up … only half of the slices from each stk seem to be loading and processing.

e.g. when I attempted to load a 60 slice stk of the blue (DAPI, in this case) channel and count nuclei, both the progress bar and the associated exported data file indicate that only 30 images are being processed. I repeated this again with a second set of stk’s I had acquired that had 100 slices in each, and this time it reads 50.

I’ve attached the pipeline (which is very basic … was just intended to test out working with the stk data type). The stk files are quite large (100+ MB), but if it is necessary to help figure out the problem, I can probably provide them via Dropbox sharing or something like that.

Any help / info / advice you may be able to provide would be greatly appreciated. Thanks! :smile:

Always,
Sean
stkPipeline.cp (4.51 KB)

Hi Sean,

This does sound odd – perhaps this issue addresses your problem? github.com/CellProfiler/CellProfiler/issues/327
I looked at your pipeline and it doesn’t have anything that stands out as odd to me, though.

You could do 2 things:
(1) Try the bleeding edge 2.0 version here (NOT 2.1). If this works then we have in the meantime fixed some bug that you are encountering.
(2) Link an example STK file by Dropbox or other means for us to take a look at.

Cheers,
David

Thanks for the follow-up, David! I’ll have access to my computer tonight, and will endeavor to follow-up with the steps you suggested to try and remediate the problem. Thanks again! :smile:

Until I have time to try the version suggestions tonight, here’s the stk file (stack of 60 10x images, Texas Red filter; cells are printed in a 3d alginate scaffold on top of a large polystyrene pillar). Let me know if the link’s not working, first time using Dropbox’s public share system.

dl.dropboxusercontent.com/u/919 … %20Red.stk

Checked the github post you referenced … I’m not the most computer savvy person, so take it with a grain of salt, but I wasn’t able to glean anything that would seem to be directly leading to the problem I’m having :frowning:

Anyway, I tried the 2.0 version you linked after uninstalling the older 2.0 version I had. Alas, same problem. I tested it with the same 60 frame stk file linked from Dropbox, and when running the analysis (using the same pipeline as before) I only got 30 images.

Any other potential ideas? Or information / logs etc. I can provide that might help? Thanks again so much for your time and assistance!

Hi Sean,

I also experience your issue of only seeing 30 images, with LoadImages set to stk. I will file a bug on this.
However if you instead choose “tif, tiff,flex, zvi movies” in LoadImages, it appears to count all 60 frames. The Image.csv indeed produced 60 rows/images. I used the new 2_0_bugfix version from our Trunk Build page which you recently installed.

Just FYI, CP is reporting a T-metadata from the stack that is 0 through 29 and a Z-metadata that is 0 and 1, so apparently our stk reader is not picking up on the Z-metadata. This is possibly a BioFormats issue, which is the image reader that we use, and/or a problem with your system in the way it writes files out, or both.

In any case, try the tif movies setting and hopefully that will work for you.
Cheers,
David

Awesome, thanks David! I will give that a try with some more test data and (fingers crossed) report back on the result :smile: