[ome-devel] OMERO/Matlab

[Michael Porter]

Hi Jonas
 
I have been using Matlab to perform various analyses on images in Omero servers for a while, and haven't come across any real limitation in the API. For anything you want to do, if there is no specific method available you can usually get the information you need by querying the DB directly.

"1. I would like to populate and query the database from within Matlab as much as possible, though I like the possibility to browse the data using OMERO. What are the limitations of the Matlab plugin?"
 
As I said above there is plenty of information/metadata available using the preset methods - once you have a pixels object in your had for any set of planes you can get a fair whack of detail by browsing the object. Anything 'unusual' can be retrieved using, for example 
"gateway.findAllByQuery(['select info from PlaneInfo as info where pixels.id = ', num2str(pixelsId), ' and theZ = 0 and theC = 0 order by deltat']);"
 
 
"2. I cannot afford data duplication at the moment. How is it possible to use OMERO without data duplication? Jean-Marie and Donald suggested that I store the file name instead of the image in the 'image' field of the database. How feasible is that?"
 
I'm not really sure what you mean by "data duplication" here. Do you have an example of the problem you foresee?
 
 
"3. It seems like the lowest hierarchy in the database are images. In my data structure, the lowest hierarchy is cells. I do have (overlapping) ROIs for each cell, and I notice that OMERO supports ROIs. Is it thus possible to have ROIs as lowest hierarchy in OMERO that can be tagged/annotated individually?"
 
Many of the analyses we're interested in also rely on ROIs. I haven't used the server-side ROI functionality (Until lately there were no ROI tables in the DB, they were all stored local to the user in xml files) but I can only see it making it easier to specify your own lowest hierarchy. Once you have all the ROIs associated with a pixels the ROIs can become your primary concern. In fact the lowest hierarchy could go down to the roishape level. Take a look at http://hudson.openmicroscopy.org.uk/job/OMERO/javadoc/slice2html/omero/api/RoiResult.html#rois for more info.
 
 
"4. To describe the data in a user-friendly manner, I would like to annotate each cell (and possibly each movie, experiment and condition) with some parameters, as well as a few graphs and possibly a collage of a few movie snapshots. I would like the parameters to be searchable, and it looks like the protocol editor would allow me to define multiple parameterName:parameterValue pairs. Is that searchable? Also, is it possible to browse through the graphs, if they're provided as, say, .png?"
 
I don't have any real experience working with the editor, but when it comes to annotations you have a few options. With the old ROI system (using XML files) you had the option of filling a text annotation for each roishape. You could get this string back and parse it for your specific parameter yourself. For more complicated annotations, like graph images for example, you can attach these to the image, dataset or project levels quite easily. I haven't done this for annotations, but uploading new images and attaching them to a specific dataset is quite straightforward. I presume it's more or less the same type of operation: upload annotation; create and populate annotation object; attach/link into dataset etc.
 
 
"5. Assume I want to run an analysis in Matlab on a subset of 'good' cells. How would I access the list of 'good' cells from Matlab?"
 
Presumably if you annotate the ROIs with some text saying 'good' you could query for these ROIs and exclude all the others. You could set this annotation either at the time of drawing the ROI, or programmatically if 'goodness' is based on some other part of the analysis workflow.
 
Hope you find this helpful! (or at least encouraging :-D)
 
Michael.
 
 
 
-- 
Michael Porter
Wellcome Trust Centre for
Gene Regulation and Expression
College of Life Sciences
University of Dundee
Dow Street
Dundee
United Kingdom
DD1 5EH

+44(0)1382 388142
m.porter at dundee.ac.uk 

The University of Dundee is a registered Scottish charity, No: SC015096


The University of Dundee is a registered Scottish charity, No: SC015096
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openmicroscopy.org.uk/pipermail/ome-devel/attachments/20091215/11f6725f/attachment.html>