[ome-devel] About SPW 'runs' (and .flex import)

[Jean-Marie Burel]

On 25/01/2012 16:32, "Anatole Chessel" <ac744 at cam.ac.uk> wrote:

>
>Hi all,
>
>    A couple of quick question about SPW: we switched to 4.3.3 recently,
>and 'runs' made their apparition in Insight, I guess corresponding to
>'PlateAcquisition' (they where not there before, right, or did I just
>not saw them?). I am not sure how that work: it seems to be a shortcut
>for a subset of well samples based on dates? So a plate have only one
>well at each position, which have a number of well sample, each with a
>position and a date. They are all equivalent as well samples, and a
>PlateAqusition links to a subset of them between two dates. In insight
>however, a plate is divided in runs and fields, which is done under the
>assumption that the same locations (different fields at the same time)
>are acquired at different time (different runs at the same position). Is
>that right?
Run corresponds to PlateAcquisition and Fields to WellSamples in the
model. Unless the name of the plateAcquisition has been set, we use the
following logic:
 * check start -end time.
 * index if nothing.
Suggestions more than welcome on how to name them in UI.
The Well is the location on the grid, a well can have X number of well
samples (I.e. an image with some location information) Those could have
been acquired at different time and physical location I.e. PositionX, and
PositionY attribute and the calculation will be done using the WellOriginX
and WellOriginY attribute of Plate.
We have made, initially, the assumptions that each well will have the same
number of fields.
This needs to be reviewed.

>
>   The thing is, in our case different runs correspond to different
>fields, for both technical and biological reasons, so 5 consecutive runs
>of a plate would be done each in a different location in wells,
>corresponding actually to 5 fields (but done plate-wise, not well-wise).
>Data (plain .flex), when imported, used to appear as different plates,
>and I had a script putting them together as different fields of the same
>plate (relinking well-sample of plate B to the wells of plate A). Now
>they appear as different runs of different plate (see screenshot
>attached); I am not sure whether that is an importation bug or an
>expected behaviour given our unexpected data and quite how to deal with
>it (as my script do not seem to be working any more -or at least insight
>is not showing any changes, I am guessing it is confused between runs
>and fields-)...
If you relink well-samples of plate B to Plate A, do you delete Plate B
after?
Do you also relink PlateAcquisition of Plate B to Plate A?
When relinking, you will probably have to modify the Index attribute of
WellSample too.


>
>   Lastly it seems that if an image in a well is tagged, and looked for
>using the search interface, there is no way of knowing where (which
>plate and screen) it comes from, as the only info is about datasets, not
>plates. I reported that a while back (it ended up here I think:
>http://trac.openmicroscopy.org.uk/ome/ticket/3871). Any plans for that?
>It would come in handy...

I will try to squeeze that in 4.4.

Thanks for your feedback
Jmarie


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