Skip to end of metadata
Go to start of metadata

Action points

ResearchIT

  • provide iBrain2 metadata to openBIS:
    • suggest the format of the file, including in particular group name, experimenter name, workflow name, workflow owner, workflow id, and process id.
    • put the file together with thumbnails dataset in a file with a fixed name (in future with raw images)
  • provide a file with microscope metadata (results.csv) together with the thumbnails
  • configure graphs so that they use right column names from the csv file with per-well analysis results
  • hide dataset tab for experiments
  • observe if the problem with loading the data has disappeared

CISD - integration tasks

  • use the microscope metadata produced by Mario to determine tile layout
  • use iBrain2 metadata to figure out in which space the plate should be created and in which project the experiment should be created.
    The data in iBrain2 should be organized using group name as the space code and user login as the project code.
  • file system access to the data:
    • discuss with Mario and Pauli if access with the linux symlinks are ok
    • ensure that Windows users do not need access to the file system after openBIS is deployed
    • an idea for the transition period: the files are saved as now and symlinks are provided to openBIS
  • iBrain2 produces now different kinds of datasets: raw images, thumbnail images, microscope metadata, per-well analysis, overlay images.
    Understand what is the order of registration (if there is any) and how can these datasets be correctly linked.

Mario:

  • feature vectors import
    • extend the file which contains focus score per well with features requested by users (like microscope focus score)
    • include a column with text values describing to which category the fear belongs (mapping of focus score ranges to human readable categories) - this would produce a heatmap which is more readable for the users
  • overlays - produce tiff files instead of *.mat files for overlays in the same directory as now

Additional new openBIS improvements

short-term

  • show a black image (or cross) whenever an image is missing for a tile
  • show on the plate layout which images have been loaded
  • when showing the magnified image the window tile should show which image it is (experiment, plate, channel, well, tile)
  • allow to modify separate channels intensity before merging them (and do the 'merged channels' transformation afterwards)
  • when choosing a dataset in the plate layout view it should be possible to see all its properties in the tooltip.
    In future a dataset will have a property which tells about the workflow which has produced it.

middle- or long-term

  • allow to compare images from a set of chosen experiments (now it's one experiment or all visible)
  • show images in analysis results table for a plate
  • choose colors for channels and save them
  • show the well image after moving mouse over the well (without a click)
  • upload control layouts (integrate with the tool which Mario and Pauli are creating for plate annotation)

Misc

  • raw data - users want to access raw images. We wait with it until full iBrain2 integration is implemented.

iBrain2 hands-on session (with Pelkmans group):

  • investigate how to plug a non-MD microscope to iBrain2 (following naming convention or implementing a new import)
  • bring back as much data from cluster computation as possible in case of a failed processing
  • ib2status web interface comments:
    • use collapsible list of items to improve page readability (for plates of a given assay + process steps)
    • display cluster job log files in case of error
    • enable custom and modular workflow registration
  • No labels