.nd2-TIFF and OMERO Conversion Errors

Hi All,
I’m seeing a weird stochastic behavior with both Bio-Formats and OMERO. I have several users who are capturing multi-XY-point time series or multi-XY series with multiple Z-positions. When we acquire the multi-XY/well series, we split the storage so that each point becomes an independent .nd2 file. However, when the user batch converts .nd2 to .tiff, some of the files get split into a series where every time point becomes a single image while the rest convert correctly. The converter macro has been working properly for months, and they need the conversion to produce consistent results as it goes into their automated analysis pipeline. We’ve also recently seen that some Z-stacks, when uploaded to OMERO will get split such that each Z-position becomes a separate file in OMERO (the behavior I usually see with multi-XY position datasets). Like the other issue, this seems stochastic as some are parsed properly while others are not. We’re using NIS-Elements 5.30.01 (Build 1541) for the multipoint t-series (we’ve since upgraded to the current build (5.30.02 Build 1545) hoping to fix this), and 5.20.02 (Build 1453) for the Z-series. Has there been a change made to how Bio-Formats parses .nd2 files? I’ve pasted in an image of the odd output file structure that we see with the .tiff conversion.

Hi Eric,

There haven’t been any recent changes in Bio-Formats that would cause such an issue but in the past there have been some reported issues regarding parsing of timepoints and positions so it may be related to one of those. Are you able to share one of the datasets showing this issue and the snippet of the macro that is performing the conversion?

I’m happy to share, but my user wants to do so through non-public means. What is your direct upload site?

Hi Eric, I can arrange for FTP access to transfer the files. If you send a PM here with an email address that I can use for credentials then I will get that setup.

I have been looking at the files uploaded and so far have struggled to reproduce the same behaviour. Do you know which version of Bio-Formats you are using with the converter macro (Can be found in Help > About Plugins > Bio-Formats Plugins)?

Hi David,
We’re having two issues, one is the tiff conversion error and the second is a problem with opening 3D multiposition RGB images. I’m running 6.6.1 on my personal laptop, and when I try to open a multi-position 1608x1608 RGB image with Bio-Formats I get what you see in the attached image.

The computer where we’re seeing errors with the tiff conversion working stochastically is running version 6.6.0. They don’t want me to run the updater as there are evidently some pieces of their code that they’re worried may break if some of the plugins are updated.

Maybe I am missing something here, but for the screenshot you posted above, is that sample file not a 2d multi position image? I don’t see any Z or T frames for it, is it supposed to have a 3rd dimension? If not I would expect it to display as separate series.

Hi David, It is 2D multi-position, but the images are totally corrupted when they are imported into OMERO or imageJ. You can notice that weird symmetrical pattern, which actually doesn’t exist in the real image.

When I import a single image it looks fine, but the multipoint images are all wrong. See attached

. I’m using the Bio-Formats importer for both.

Yeah, this appears to be a different issue, I believe it may be related to the bits per pixel being incorrect and have opened a GitHub Issue: ND2: Incorrect bits per pixel · Issue #3667 · ome/bioformats · GitHub

Is the single image captured using the same camera and software or are there any differences in the acquisition setup?

The two images were captured with the same camera and same optical configuration. The monochrome images from the Qi2 are always handled perfectly.