Strange LUT behaviour from Micro-Manager

[ratamero]

Hi there,

we have one of our PIs who's using Micro-Manager for their microscope, saving everything as OME-TIFF stacks with metadata companion files. Their images are usually DIC + FITC two-channel; on the ImageJ instance inside Micro-Manager, these show up correctly as greyscale and green, respectively.After importing into OMERO they show up as red and green, so I decided to investigate whether Bio-Formats was missing any colour information and just defaulting the DIC channel to red in the absence of relevant info. Opening these images via Bio-Formats in Fiji shows these channels as greyscale and magenta, though, so there was no consistency there. OMERO and Fiji, both using Bio-Formats, managed to get the colours wrong in two different ways! Both of them are using the same (newest) Bio-Formats version.



These are the first few lines of the metadata XML file as well:

Code: Select all

<OME xmlns="http://www.openmicroscopy.org/Schemas/OME/2016-06" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://www.openmicroscopy.org/Schemas/OME/2016-06 http://www.openmicroscopy.org/Schemas/OME/2016-06/ome.xsd">
<Instrument ID="Microscope">
<Detector ID="Ratiga R6" Manufacturer="PMUSBCam00" Offset="400.0"/>
<Image ID="Image:0" Name="RW10006_1_MMStack">
<AcquisitionDate>2018-10-29T11:46:13</AcquisitionDate>
<Description/>
<InstrumentRef ID="Microscope"/>
<StageLabel Name="Pos0" X="-448.0" XUnit="µm" Y="4307.0" YUnit="µm"/>
<Pixels BigEndian="false" DimensionOrder="XYCZT" ID="Pixels:0" Interleaved="false" PhysicalSizeX="0.09014424" PhysicalSizeXUnit="µm" PhysicalSizeY="0.09014424" PhysicalSizeYUnit="µm" SignificantBits="16" SizeC="2" SizeT="439" SizeX="650" SizeY="650" SizeZ="1" TimeIncrement="650.0" TimeIncrementUnit="ms" Type="uint16">
<Channel Color="-1" ID="Channel:0:0" Name="DIC" SamplesPerPixel="1">
<LightPath/>
<Channel Color="-16724992" ID="Channel:0:1" Name="FITC" SamplesPerPixel="1">
<LightPath/>

Any ideas/guidance on why that is happening would be very helpful!

Best regards

Erick

[mtbc]

Dear Erick,

Thank you very much for all the information, it's helpful. Could you submit an example data set please for us to experiment with? http://qa.openmicroscopy.org.uk/qa/upload/ if it's not too large. Also, which version of OMERO did you use?

The OME-XML you quoted defines the first channel as white and the second as a reddish magenta. (The color numbers are 32-bit signed RGBA values.) I'm certainly curious to see how any of the channels appear green!

Cheers,
Mark

[ratamero]

We're running OMERO 5.4.9 on our server. I'll get a sample dataset from the PI and upload is as soon as I can! Thanks for the prompt response.

edit: file uploaded!

[dgault]

Hi Erick, thank you for providing the sample files, we were able to begin testing them today. From the OME-XML metadata available it looks as though the colours from the FIJI import match those described in the file.

What I have not yet figured out is why the colours exported in the XML differ from those being displayed in Micro-Manager. We will continue to investigate this in the next few days and update this thread once we have more information.

David Gault

[dgault]

As an update to this, there are 2 separate issues here. The first is that the color values stored in the metadata txt and the OME-XML do not match the Green you expected when exported via MicroManager. This is something which the MicroManager team will need to fix.

The second issue is that the behaviour of Bio-Formats differs depending on if you select the ome.tiff file or the metadata file. When selecting the ome.tiff file Bio-Formats is using the OMETiffReader and is correctly populating the channel colors from the OME-XML. When selecting the metadata file the MicroManagerReader is being used and this is not populating the colors for the channels resulting in the default Green/Red which you see in OMERO.

I have opened a Trello card to track this problem https://trello.com/c/QhDWG4aL/94-microm ... nel-colors and I have also opened a PR to put in place a fix for the second issue mentioned: https://github.com/openmicroscopy/bioformats/pull/3285

[ratamero]

Brilliant, thanks a lot David!