"Strange" behaviour from Fiji,Bio-Formats importer. with MM.

[i.munro]

Good Afternoon

I wonder if you can help me.
I'm attempting to use the Fiji Bio-Formats importer's "Specify range for each series" option to read
a subset of frames from time-series data files generated by Micro-Manager (MM)
When the data is larger that 4G MM generates a fileset of multiple files.
Superficially all seems well and the Importer handles the fileset well.

However I have noticed that I can still open (say) frame 9999 from a 10,000 frame series even after
I delete the 2nd file of the set.

Checking with the bio-formats showinf tool "./showinf -nopix" also seems to give no errors after deleting
the 2nd file in the set.
Can you tell me if this is expected behaviour - it seems counterintuitive.

The data was taken with v1.4 of MM and I have updated Fiji (manually) to use the recent 5.2 version
of the Bio-Formats plugin.

Many thanks

Ian

[sbesson]

Hi Ian,

thanks for the question and the uploaded fileset. I was able to reproduce a minimal example highlighting the issue using one of our sample filesets.

The zip attached to this forum post contains a modified version of the files linked above with the OME-XML header of the second OME-TIFF modified to remove the `UUID` element of the root `OME` attribute

Code: Select all

<OME xmlns="http://www.openmicroscopy.org/Schemas/OME/2016-06" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" Creator="OME Bio-Formats 5.2.0-m4" xsi:schemaLocation="http://www.openmicroscopy.org/Schemas/OME/2016-06 http://www.openmicroscopy.org/Schemas/OME/2016-06/ome.xsd">


If the tubhiswt_C1.ome.tif file is removed from the fileset, then the reader initialization using e.g. showinf will fail in the setId call with an Unmatched UUID exception.

However if the tubhiswht_C0.ome.tif file is removed from the fileset, the reader initialization does not throw an error and the planes which should have been read from the missing file are also read from tubhiswht_C0.ome.tif.

There are a two orthogonal issues here:

• the OME-TIFF reader should make a best effort to behave similarly in both cases and ideally fail at initialization unless we allow a -strict/-loose flag


• in terms of specification, it might make sense to make the root OME UUID attribute required or minimally recommended in the case of multi-file OME-TIFF filesets. This recommendation only applies to companion filesets at the moment.

Best,
Sebastien

[i.munro]

Many thanks Sebastien. Is there any way I can add the missing UUID using Micro-Manager for future acquisitions?

[sbesson]

Hi Ian,

sorry for the delay. First of all, after a quick check, the multi-file OME-TIFF should open as expected if all the files are present. The issue described above is only valid for partial filesets.

If you want to add the UUID information to an existing OME-TIFF fileset, you can certainly use tiffcomment as a way to edit or replace the OME-XML.

To do it in Micro-Manager itself, it might be worth engaging the conversation with their development team using either their mailing lists or their source code repository to enforce the top-level UUID is filled when saving images as stack files especially in the case of filesets spread over multiple files.

Best,
Sebastien

[i.munro]

Many thanks Sebastien.

Ian