2016-10-11 Tuesday Team Meeting
Dundee: Will, Petr, Dominik, Balaji, Jean-Marie, Gus, Mark, June, Roger, Simon, Helen, Remote: Sebastien, Jason, Josh, Andreas, Jerry, Emil, Chris, Eleanor, JB, Liza, Melissa, Ola,
Presentation by Jia-Ren Lin (Jerry) from Harvard (Sorger Lab) - "Probing tumor heterogeneity with Cyclic Immunofluorescence (CycIF), a robust, multiplexed imaging method"
Jay Copeland also present.
Different type of image - need help to manage data
Drug response study - lots of heterogeneity seen using microarray of tissue samples.
Existing techniques limited:
- Mass Cytometry slow & low resolution, Quantom dots limited channels.
- Quantum dots PE, MultiOmyx (chemical bleach) proprietary
Cyclic ImmoFluorescence (CycIF) - single cell multiplexed measurement with cyclic IF
Fixed cells -> 1st IF -> Flour inactivation -> 2nd IF -> Flour inactivation
Allows to get multiple probes per channel
In live cells can inactivate flour with H202 or pepsin (cleaves Abs)
Many channels/probes for single cell allows cross-correlation
96-well plates (use 60 wells) 9 rounds of CycIF
3 channels + Hoechst stain per round => 27 channels (Hoechst DNA stain for registration)
Import images into OMERO - many channels in web image viewer!
Analyse individual cells in images
PhenoGraph & Metacluster, Wanderlust algorithm etc.
viSNE for visualization of high-dimensional data
Can be used to disect and classify drug responses
Use High-Res (60x confocal InCell 6000) allows analysis of morphological features (not just intensity)
Quantify and cluster morphological features (heatmaps etc)
Live cell imaging - use cell-cycle fluorescent markers to probe cell cycle phase
RNA FISH CycIF
Summary:
- High content, high resolution, upto 10 cycles in each cell
- 140 commercial antibodies
- CycIF for FFPE tissue/slides
- Pre-stain with 2ndary Abs (use image as background) => Flour inactivation => 1+2 Abs => inact => conjugation
- CycIF cycles 143 fields, tiled to give Big image, 8 cycles - many channels
- Image segmentation and analysis of cells
Tissue microarray
In OMERO, each image in Dataset for each tisse microarray sample
Collaboration is important - share data with pathologists etc.
Need better image annotations
Summary:
- CycIF can be used in both liquid bopsy & FFPE
- Can improve traditional IF in FFPE (control autofluorescence & 2ndary background)
- http://lincs.hms.harvard.edu/lin-NatCommun-2015/
- http://lincs-omero.hms.harvard.edu/webclient/userdata/?experimenter=953
Will: How to improve web image viewing?
Over 4 channels hard to display as RGB
Need to be easier to choose 4 channels in web viewer - E.g. assign channel to Red, another to Green & Blue etc.
OR calculate summary images
Current web viewer - channel names not readable on buttons
Too many channels - not scrollable in left of viewer
use Z to store each round of cycIF - but can’t name individual channels then
Project Timelines (2-3 minutes each)
Spaces
-
Files/Formats/OME Files (Sebastien)
- 5 release tomorrow
- 5.3 work with Roger on builds etc
-
Metadata (Josh)
- New version demo3 with mapr in preparation
- Jupyter notebooks for IDR - simple analysis examples
- Copy of IDR available to download via rsync
-
OMERO 5.3.0
- SPW display in progress
- Web apps package renaming ongoing - OMERO.figure nearly done. Push to Pypi
-
Outreach
- Balaji presenting at PiCLs
- GRE symposium
- Balaji videos in production - Matlab toolbox
-
Files/Formats/OME Files (Sebastien)
Other releases/upgrades:
-
Figure
- Waiting for renaming to complete before restarting work
- ImageJ
- Learning
- Sysadmin
-
Website redevelopment/redesign (Liza)
- Recent changes to BF and OME files sections. Now working on Products/omero
-
Smuggler (Montpellier)
- Met with Icy team to discuss OMERO connection
-
Figure
Glencoe Update (Chris)
- jpeg XR work, setup on various different platforms.
- Finalising API - doing everything in memory
- Preparing to port over to open source
- Test Datasets will be publicly available
AOB (5 mins max - technical discussions should be highlighted to relevant people and rescheduled)
- PR merging strategy discussion
-
Improve PR Opening:
- initiate discussion with relevant people, cc them in description etc
- List tests added etc.
- List as much info as possible to help reviewers
-
Improve review:
- Functional review
- Code review
- Everything is public - schedule zoom discussion if needed
- New Review feature
-
Improve Merge:
- Need clear statement of ‘Good to merge’
- Bigger PRs need more reviewers to validate it
- If mergeable and not merged, bring to attention. E.g. list at standup.
- If not mergeable, and inactive, Close until ready to review again