2018-12-11

Dundee: N/A

Remote: Simon, Frances, Kenny, Mark, Dom, Riad, Jason, Josh, Liza, Chris, June, Seb, Melissa, Andreas, Wilma, David,

Start: 2:00 pm

1. Accepting minutes from last meeting

2. Project Status

(2-3 minutes each)

• Data-In

  • Seb: Bio-Formats 6 - upcoming freeze steps

    • JRS: NIH directors etc. interested in the #industry linkage

    • Milestone release before Christmas

  • SpaceTx (Josh): finally getting users for the fov-writer. See

    also roadmap in #hca.

    • May need new PRs in final Bio-Formats 6
  • Seb: Iviewer performance/usability/bug testing this week on

    existing and new IDR datasets

• Data-Out

  • Mark: OMERO.downloader: Pushed commits hiding the broken export

    options. Drafted introductory blog post. Plan to chat to Jean-Marie about what more needs done to release.

  • Riad: From Gradle to grave for OMERO builds, should fix problems

    with build dependencies

• Glencoe

  • More microservice work, including a vertex inside OMERO using

    Hazelcast

  • Linking microservices with PostgreSQL session store, would also

    allow Django to store sessions in PostgreSQL as an alternative to Redis

  • See open PRs

  • More Bio-Formats work

• Community

  • I2K Below.

  • JRS: [ASCB

    card</u>](https://trello.com/c/bQjVIQlS/226-ascb-embo-2018-7-9-dec-2018) e.g. BDV test data

1. AOB

(5 min. max; tech. Discussion should be highlighted to relevant people and rescheduled)

1. Main Topic - I2K Feedback (card)

(20-25 minutes plus 15 minutes questions max)

• Bio-Formats

  • +1 for TIFF-Pyramid ([see blog

    post</u>](http://blog.openmicroscopy.org/file-formats/community/2018/11/29/ometiffpyramid/))

  • Then: HDF5 followed closely by

    N5 (tree of JSON files with pointers to tensors chunked into smaller files, optimised for parallel compute including writing) but N5 is still insufficient. Upcoming discussions on migrating BDV to N5.

  • image.sc shared forums

    including Bio-Formats

  • IO: APR (Adaptive particle representation) (or not), next-tiff

    (“performance”)

    • https://www.nature.com/articles/s41467-018-07390-9

      (variant of HDF5)

  • Melissa: User request for Docker bfconvert

• EOSC anyone?

  • Jean-Karim promoting it; Galaxy / KNIME

  • JRS: EOSC is a virtual cloud of clouds…

  • BIA is part of EOSC, officially approved by EMBL, will be built

    by EBI

• OMERO

  • Firewall issue during workshop, ports 4063/4064 blocked.

    Something worth anticipating for future outreach sessions (Web port over SSL)?

  • CATMAID, may want to demonstrate integration with OMERO for

    electron microscopy

  • KNIME some interest in building a workflow

• Analysis

  • Deep everything: Broad, Ilastik, lots of u-net, Keras: yes,

    Theano: no.

    • New tools for dealing with the DL blackbox aspect

      (visualization and validation)

    • Lots of different strategies for ground truth i.e. not only

      manual labelling, secondary channels etc

    • See trello card for links. We should try something.
  • Clear tools for SMLM (Single-Molecule Localization Microscopy)

    and tissues (also clear data)

  • BDV Ecosystem e.g. including tracking (Mastodon/Trackmate3),

    segmentation, stitching, visualization tools working with BDV

    • TrackMate 3 tested with 10M points.

  • Interest in having ground truth and labels in IDR.

    • Question of where to store the models.

• Misc

  • Funding & “Research Software Engineering”

  • Discussion on management of RDF and unstructured metadata (see

    image.sc)

  • Conda

  • Lots of users still relying on ImageJ1 Macros

• Next meeting: 2020 in Madison