Ziess bioformats cannot find multiple angles for import into MVRApp

[HHathrell]

I am trying to import a dataset through Ziess Lightsheet Z.1 Dataset (LOCI Bioformats), but it can't seem to recognise there is more than one angle present. The dataset is saved using the names ...G1, ...G1(1), ...G1(2) and then ...G2, ...G2(1), ...G2(2), and ...G3, ...G3(1), ...G3(2). This is standard naming of angles 1,2,3 by the Zen software. The files are all in the same folder, and it is correctly reading the metadata to find that each file contains both left and right illumination views. I am using the first ...G1 file (or G2, G3), but whichever angle I give it, it does not find the others.
I also don't know of a way to import more data after creating the original xml, so I could add each angle separately?
Any thoughts would be appreciated.

[StephanPreibisch]

Hi, how do you define the dataset? Could you explain it in detail please so I can advise?

[HHathrell]

The zen programme adds the "G1" "G2" etc to the name of the file automatically as it acquires each view angle - there is no choice in changing this when you acquire the data set. The dataset I am trying to register has angles '1','2' and '3' taken at 0,120 and 240 degrees respectively (by rotating the sample in the Z.1). Each file contains two separate left and right illumination directions also - these are being recognised by the App.

The file names are (timepoint in brackets) e.g.
TL-3views_G1.czi
TL-3views_G1(1).czi
TL-3views_G1(2).czi

TL-3views_G2.czi
TL-3views_G2(1).czi
TL-3views_G2(2).czi and the same for ...G3...

All the files are in the same folder.
I have been trying to define the xml using the Ziess Loci Bioformats option, with:
by using the first 'TL-3views_G1.czi' file. However it is only finding the G1 set, not G2 or G3: 'Angles (1 present). I have tried starting with the (1) time point files and the G2 and G3 equivalents, but no luck either.

[StephanPreibisch]

Hi, I assume something is not right when you acquire the datasets. By, the way, you can change how the Zeiss software saves the files somewhere in the acquisition panel, I forgot where, but you can. This, however, rather looks like you pressed "acquire" multiple times for this dataset as the header-file (the one without brackets) is normally created only once per acquisition - that can consist of many timepoints, channels, angles, and illumination directions. So the reader will not recognise multiple, independent acquisitions as you seem to have here. Maybe the software changed, but we have one of the newest versions on our microscope as well, and this is also how I know it.

Also, your description does not fit the file name structure. If each angle would have two illumination directions I would expect
TL-3views_G1.czi
TL-3views_G1(1).czi

TL-3views_G2.czi
TL-3views_G2(1).czi

TL-3views_G3.czi
TL-3views_G3(1).czi

If you want to can send me these files and I can have a look.

All the best,
Stephan

[HHathrell]

Matt says:
'Perhaps there is a differences in the software - we have the 2014 update.

I don't understand the comment about the pressing the acquire more than once. I set up the 3 views and start the acquisition and all 3 view/groups are saved as you see them.

Puzzling....'

The files are over 3 GB each, do you know a way I can send them?

[HHathrell]

We also have a ...G1.queue file (and same for G2 and G3) I have not tried yet?

[StephanPreibisch]

I am not sure what you mean with "groups" though. I did not use anything like that, maybe that is the reason. What are they good for?

[HHathrell]

Hmm... we will have a look at the Zen software for now and we may have to resave the data, thanks Stephan.

[StephanPreibisch]

Hi, you can also import them one by one, resave as TIFF, and then define a new dataset based on the TIFF's ...

[StephanJanosch]

I am not sure what you mean with "groups" though.

I once told you @StephanPreibisch, that these groups should correspond to your tiles. If you image a worm, then the top section is group one and the bottom section is group two. Here is an example:

5 angles, two groups. Check the y position and you see, what the top section is and the bottom section. Support of these mvl files would have been great.

150618_TRG_1446_Worm1_DAPI.mvl.txt

[StephanPreibisch]

We do the same kind of imaging (multiple Tiles) since quite some time without using groups in Zen, so it is definitely not necessary for that - metadata allows to separate angles from tiles ... maybe we are setting it up in an easier way than it is supposed to be. Anyhow, we are working on fully supporting tiles, but it is not trivial for many reasons.

However, your comment mixes two different issues. (1) The import of such a "split"/group dataset, which is at the moment possible through the mentioned workaround of saving TIFFs first. (2) Processing of tiled, multiview data. There will be a release soon though supporting (2).

Regarding (1): Grouping seems to be an undocumented/unsupported feature as LOCI Bioformats does not seem to recognise groups that belong together. It should be implemented on the LOCI Bioformats side, not inside the Multiview Reconstruction as a workaround that will fail once Bioformats supports it.

I am very sorry if I missed your comment about Zen groups @StephanJanosch ...

All the best,
Stephan