Import of images of Evotec/Perkin Elmer flex format

[frank]

Hi

Today I made my first experiences with the upload of HCS images to OMERO. I tried to import 324 images of the Evotec/Perkin Elmer format FLEX that are located in one directory. The names to the files are like 002001001.flex (ca. 1.3 MB), 002001002.flex, etc. I thought that one flex file stands for one well of a plate.

The importer "says": "108/108 ; 1/1". What does 108 in this case mean?

After the import I can see in the "scrrens" tab that there are a lot of screens with the name of the folder that is one hierarchy higher than the folder with the flex files. Bellow each screen there is a "run".

If I click on one "run", parts of the plate (6*9 wells) are shown together with the 7*9 images (Z*T*C=1*1*1) and it seems to me that these 6*9 images belong to one flex file.

At the moment I am a bit confused. Can anyone help? Is there a documentation of the flex format?

Best Regards, Frank

[jmoore]

Hi Frank,

which version of OMERO did you import this with?

Cheers,
~Josh

[frank]

I use OMERO version 4.4.10.

Frank

[rleigh]

Hi
I tried to import 324 images of the Evotec/Perkin Elmer format FLEX that are located in one directory. The names to the files are like 002001001.flex (ca. 1.3 MB), 002001002.flex, etc. I thought that one flex file stands for one well of a plate.
The importer "says": "108/108 ; 1/1". What does 108 in this case mean?

Looking at the FlexReader class (https://github.com/openmicroscopy/biofo ... eader.java), you are correct that there is one .flex file per well (and field).

The number here refers to the number of planes, rather than wells. For example, if you have a single well containing a timeseries and/or z-stack and/or multiple fields, these will all contribute to the plane count.

In addition to this, when you import a single .flex file, this will also import the rest of the plate--selecting any file in a plate is effectively going to import the plate. The collection of files which make up the plate are treated conceptually as a single entity.

After the import I can see in the "scrrens" tab that there are a lot of screens with the name of the folder that is one hierarchy higher than the folder with the flex files. Bellow each screen there is a "run".

If I click on one "run", parts of the plate (6*9 wells) are shown together with the 7*9 images (Z*T*C=1*1*1) and it seems to me that these 6*9 images belong to one flex file.

When you performed the original import, you would normally at this point select the "screen" to put the import into (though this is optional). OMERO organises the data hierachially in the order screen, plate, [run], well. So you should have one (or more) plates under the screen for the flex file you imported after the import completes. I'm a little uncertain of the utility of the "run" myself, but it looks like it's probably intended for reimaging of the same plate. One run per plate is typical. Selecting the run with a double-click will show you an overview of the plate, from which you can then select wells.

It certainly shouldn't be the case that the contents of one well should appear in other wells. If it looks like something wrong is happening here on import, you can send us the data if you would like us to investigate this in more detail. If you zip up the entire folder, you can upload it here: https://www.openmicroscopy.org/qa2/qa/upload/

Kind regards,
Roger

[frank]

Hi

Unfortunately it is not possible to upload these images because these are part of an active project. I am sorry! These images were generated by a Perkin Elmer (PE) Opera system. Maybe one from PE can help?

In the directory there are 324 flex files. If I upload e,g, the file 002001001.flex then a plate layout 6*9=54 is shown. One can toggle between "Field1" and "Field2" which changes also the channel name, i.e. 108 images.

I could verify that the file name codes the well. I will manipulate the original file name like 002001001.flex to 200-100-1 so that it is easier to explain. 200-100-1 stand for well B1 and holds 2 images (fields?, channels?). 00-700-900-1 stands for well G9, i.e. 2nd number=row, 3rd number=column of plate.

If I compare 00-700-900-1 with 00-700-900-6 they are the same well, but hold dirrent images.

And: The plate layout is shown as a 384 well plate, whereas a click on an image mentions a 96 well plate in the in right hand pane.

Frank

[mlinkert]

Each .flex file does contain one field, i.e. part of a well, but if there are gaps in naming or inconsistent sizes between files then there can be missing wells or fields. Bio-Formats/OMERO does the file name grouping on a "best effort" basis unless the original .res or .mea metadata files are present - if you have access to those files, copying them into the same directory as the .flex files may help.

If you aren't able to send the files themselves, would you be able to send a long directory listing (i.e. 'ls -hl' on Linux or Mac OS X; 'dir' on Windows) and maybe a screenshot of the imported plate in OMERO with thumbnail images blurred out? That would help us to better understand what might be going wrong. Do you also know what the expected plate dimensions (number of wells per row and column), field count, and channel count are? It's not entirely clear to me if 6 * 9 wells is the expected plate size, or if a 96 or 384 well plate is expected instead.

[frank]

Hi

We expect a 96 well plate with images for 54 wells (B2-G9). That is correct in OMERO. Furthermore there are two images, i.e. 2 channel) in one flex file. You can toggle between these two channel by using the drop down list ("Filed#1", "Field#2") in the middle pane. In the middle of the right pane one can see the channel name. And we expect 6 fields per well. That is not shown in Omero. See the attached png file.

flex-format.png (134.74 KiB) Viewed 3411 times

If I import the dataset of 324 flex files together with the mea and res file in the same directory, OMERO uploads 108 images (54 wells * 2 channels), but not 54 wells*6 field*2channels=324 flex files * 2 channels.

The name structure of the 324 flex files is very easy:
002001001.flex, 002001002.flex,002001003.flex,002001004.flex,002001005.flex,002001006.flex,
002002001.flex, 002002002.flex,002002003.flex,002002004.flex,002002005.flex,002002006.flex,
.....
.....
007009001.flex, 007009002.flex,007009003.flex,007009004.flex,007009005.flex,007009006.flex


Hope that help to find the answer,
Frank

[jmoore]

Hi Frank,

sorry for not having gotten back sooner. Without sample files, we haven't been able to reproduce what you're seeing. If there's anything else we can do to help, please let us know.

Cheers,
~Josh.