2018-08-21

Dundee: Will, Petr, Kenny, Dominik, Jean-Marie, Mark, Simon, Sebastien, Mike-Wicks (EDI), June,

Jason,

Remote: Josh, Wilma, Emil, Melissa, David

Start: 2:00 pm

1. Accepting minutes from last meeting

2. Project Status

(2-3 minutes each)

• Summer epics

  • Import (Josh)

    • Merging ongoing, into develop.

    • Existing work: logs, metrics, Insight GUI PR

    • Upcoming: no default thumbnails, parallel import CLI

    • Few other miscellaneous PRs (channel color, etc)

    • 5.4.8 RC build targeted for 31st of August (Friday)

    • Various testing necessary

    • Careful to test all Insight & BF changes

    • JM: Help docs screenshots need updating (useful for

      workshops, esp if no time to demo import)

      • Basel: doing import (don’t need help)

      • Sydney: no import → need help

      • Okinawa: probably not

      • Kobe: perhaps

    • Jason: thanks for summer work on import

    • Discussed holding back release for more import work.

    • JM: Some docs PRs

  • IDR (Dom)

    • Prod51 set up by Seb

    • Idr0041 currently being reannotated on prod51 (Seb)

    • Idr0043 (HPA) - 15TB, 500,000 images: Still trying to import

      on idr-testing. Now uses %hash% instead of %timestamp% for directory names to avoid clashes when importing in parallel. Import with “skip all”. But still latest import attempt just failed again.

    • OMERO might cope better with images in Plates? Single

      fileset? etc.

    • Idr0045: Going to import on prod51 when idr0041 reannotation

      is finished. Still waiting for genotype annotation from J-K, but can be added later in case it’s too late for idr 0.5.1 release.

  • Mail (Kenny): New mailman server:

    • Testing mailing lists with spam filtering. Reply/Reply-all

      behaviour changed. Seems to be working OK over last week.

    • Mailman upgrade would require new server, but not

      necessarily a bad thing.

    • Approx 2-weeks’ work.

    • One option: Use google mail. Or outsource to Mailman

      provider

      • Jason: Longterm stable provider? Archiving?

      • Prefer ome ‘branding’ instead of google.ome etc.

      • Discourse? (IJ etc)

      • Josh: Mailman is ‘easiest’ (what we know how to do).

        Usability vv google

      • Could investigate cost of moving to Discourse?
    • Kenny: will look at branding of Google services, and our own

      Mailman

  • Training:

    • New public alternative jupyterhub server.

    • Backup server since Jupyter notebooks are critical to

      workshops

    • New www.ome/training pages and ome/apps page

    • Jason: ITR (Eurobioimaging deliverable) is nice bridge

      between IDR and training work. JM: recently added Trakmate

• OMERO 5.4.8 timeline (see import above)

  • Release early-mid September

• Glencoe

  • Pyramids work ongoing

  • OMERO-metadata PR open. Needs discussion & test fixes

  • Melissa: Channels ome PR

  • Chris (away): OMERO.parade PR

• Community

  • #hca (Josh): still waiting on real data.

    • Will go in _from_hca directory.

    • While waiting can work on validation, then time to ramp up

      the S3 work.

    • Reading TIFFs from S3. Various options re: performance.

  • #ome-ssbd is quiet, but will start ramping up

  • Prep for Sydney, Japan, Cambridge… ~1 a month

1. AOB

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

1. Main Topic

(20-25 minutes plus 15 minutes questions max)

• Mike Wicks - EDU

• Comparative Pathology Workbench (CPW)

• HCA (Human Cell Atlas) CZI-funded project, 6-8 Sub Areas

• CPW: Scienc-y bit (conceptual coordinate model of the small and

  large intestine (not discussed today)

• Who? Mark Arends (PI), Richard Baldock (consultant, models & tools),

  Albert Berger. (mapping model, bioinformatics).

• Objectives:

  • Prototype for pathologists

  • To Visualise, Arrange, “Annotate”, Analyse their data

  • Share dta with other pathologists

• Grid/Matrix presentation of Data

• Before: rows “meant” something. E.g. Mouse / histopathology etc.

• Now: Completely freeform grid - anything goes anywhere - user can

  choos.

• Functionality of CPW (typical Pathologist’s workflow)

  • Biuild a matrix or matrices

  • Assign data to cells - 1 image per cell

  • Share matrix

  • Collaborative comments

• What Data is in a CPW cell? Mostly Images, hence OMERO.

  • Histopathology

    • Image - Microscope

  • Radiology

    • Image - X-ray. Format?

  • Genomic

    • From Ensembl

  • Transcriptomic

    • RnaSeq - e.g. Heatmaps

• Architecture:

  • SPA web application

    • Database back-end to store Matrices, Django

    • Annotate matrices / cells

  • OMERO:

    • Authentication

    • Host Images

    • ROIs

  • Blogging Engine? - collaborative commentary

• Free-form grid metaphor.

  • OMERO flattened tree

  • Of interest to OME..?

• Integration with OMERO

  • Authentication

  • Annotation

  • Avoid Reinventing the wheel

• Collaborative Opportunities:

  • New metaphor for organising OMERO data?

• SPA with React - NO! Django more suitable

  • Demo built in Django