Ilastik not opening .tiff files coming from ICY

Hello

I have an imaging mass cytrometry dataset which I’m trying to analyse using ICY and Ilastik. So far I have used ICY for noise reduction, 2x scaling and random cropping ( more information about these steps in this link)

I’m trying to use Ilastik now for pixel classification but when I load the .tiff files produced with the ICY plugin random-crops, Ilastik just remain inactive for hours (picture and tiff sample attached).

Any idea of where the problem is?

Regards
Juan

ICYcrops.zip (3.4 MB)

1 Like

Hi @Juan_Ojeda,

sorry that you have problems with this data. First of all it would be interesting to know which version of ilastik you are using. I have tried 1.3.3post3 and could open the data.
However. I have the feeling that ilastik might still not recognize it properly. My feeling, when looking at the data is that there might be multiple 2-channel images in one tiff file? Is this a time series? For ilastik it would be better to export those time frames individually (so multiple 2-channel images). I have mentioned Marion from ICY here, as she might know right away how to export such data in order to use it in ilastik.

Cheers
Dominik

cc @MarionLouveaux

2 Likes

Hi @k-dominik, @Juan_Ojeda,

At least what i can say is that Icy always saves TIF images as OME-TIFF and the image generated is indeed a bit special as when i try to open it i can see internal data has double type and the image is composed of 22 channels which might be an issue for Ilastik. I think it would have been better to stack the image in another dimension (T) to make it easier to open on others softwares but maybe @Juan_Ojeda did that on purpose.

Best,

– Stephane

2 Likes

Hi @k-dominik and @Stephane

Thank you so much for your help.

k-dominik, I use Ilastik 1.3.3post2, I will download and try post3 tonight (travelling at the moment with just basic internet). The tiff is from a imaging mass cytometry data set where we measured 22 channels independently and then made a stack. There is an explanation of how I made it in the link of my initial question.

Stephane, I’m happy to try any suggestion but I don’t know how to stack the image in another dimension and I’ll need some help. My knowledge on this field is very limited, I (with the help of amazing people like Marion and everyone else who has shared some advice) am just trying to build from scratch a pipeline using ICY-Ilastik for imaging mass cytometry data analysis.

Regards
Juan

1 Like

Hi @Juan_Ojeda,

Unfortunately this is not an operation we can do easily using the integrated commands in Icy’s GUI so you have to pass through a small protocol to do that task…
I made the protocol for you so you can just drop it in Icy and apply on your image, it will transform your channels image to a timelaps image :
Channel to Time.protocol (4.6 KB)

Hope that helps :slight_smile:

Best,

– Stephane

1 Like

Hi @k-dominik and @Stephane

I downloaded 1.3.3post3 but I couldn’t load the file. No idea why I can’t open the data and k-dominik can. I just created a pixel classification project and loaded the tiff. What am I doing wrong?

I also tried to load Stephane’s protocol in ICY (thank you so much for taking the time to do this!) but again I just got an error message (picture attached). Should I post a new question about this error or is there any really evident mistake in the picture?

Sorry about the lack of good news and thank you for your help.

Regards
Juan

Hey Juan,

So I definitely saw those 22 channels in ilastik. So each channel shows the same spacial location? Wow :slight_smile:

1 Like

okay that’s weird. I’ll try on a windows machine later (I’m on linux) and see if I can see the same problem there.
In the meantime, what will always work is loading the data via Fiji and exporting it with our import export plugin to hdf5, then loading it into ilastik.

1 Like

Hi k-dominik

Yes, you’re right. There are 22 markers in the same slide, so it’s the same spatial location in every channel. Channels 1 to 20 show markers associated to cellular cytoplasm and membrane. The last two channels show markers associated to DNA so I use them to identify the nuclei of cells.

I thought about import export plugin as last option because ICY preference for tiff files but I will use it if there is no other option.

Regards
Juan

Hi Juan,

Protocols plugin has been updated some weeks ago. Just be sure that your Icy is up to date and it should open correctly (you can force update by searching for “check update” in the search bar).
That should fix your issue :slight_smile:

– Stephane

1 Like

just to let you know @Juan_Ojeda, I can see the problem on windows… Not sure about the cause though, need to look into this.

1 Like