Image Informatics for Biological Microscopy
Copyright © 2005 -- All Rights Reserved
Chris Allan
Division of Gene Regulation and Expression
Wellcome Trust Biocentre
University of Dundee
Scotland
Microtubules
DNA
Live cell imaging- shRNA screen of unknowns
Scram
LRR
hDam1
Links Between Mitosis & Neurogenesis
Chick Embryo Neural Tube
GFP-tubulin
3D Stack/7 mins
38 hours total
5GB
X 4:8
512 x 512 x 30 x 342
Arwen Wilcock
Links Between Mitosis & Neurogenesis
Arwen Wilcock
Kate Storey
HCS for Small Molecule Discovery
Nuclear Import Screen: Yan Feng, Harvard ICCB
Light Microscopy at Dundee: Dealing with Data
5 Multi-D WFMs
(+488, 532, 410 nm lasers)
3 Multi-D LSCMs
1 MP LM
(+lifetime)
Data Store Size (Tb)
Light Microscopy at Dundee: Dealing with Data
12TB
6TB
• 9228 CDs
• 1272 DVDs
• 18456 CDs
• 2544 DVDs
Current Paradigm
No Standards
Metadata
Digital
Image
Acquisition
System
Raw
Image
Metadata
Transformation
Processed
Image
Links between
Experiment?
Image?
Analysis?
Visualization
Quantitative Analysis
“Result”
http://www.openmicroscopy.org
Towards Image Informatics: OME
Digital
Image
Acquisition
System
Metadata
Raw
Image
Quantitative Analysis
OME
Complex Data
Management
Metadata
Visualization
Processed
Image
http://openmicroscopy.org
Open Microscopy Environment (OME)
•Open system that specifies a data model for biological
microscopy
•An XML file for data migration and transfer
•Software tools that store images, metadata, and analytic
results in one or many relational or file system based
databases
•Designed for data model extensibility
•Open source
•Community driven
http://www.openmicroscopy.org
OME: Building a Community
What is the minimal set of metadata we need to
record about an image?
User 1: ”Data scaling (x, y, z, t, channel)
Microscope type
Excitation and/or Emission wavelengths
Pinhole (if appropriate)
Collected by X
Date collected
"Raw" or "Processed"
User 2: “None. However important metadata might be, images still can
have enormous value without metadata.”
http://openmicroscopy.org
OME: Building a Community
What actions do we take if an incoming file does not provide some of this
metadata? Do we use defaults, or do we force the user to give us the
data?
User 1: ”Defaults will be a disaster - people or too busy or too lazy - force by any means!
Offer batch/templates for meta data annotation and if there must be some
compromise then a grace period of say 30 days may be issued after which the
images are removed (uploader could be spammed during the 30 days for
reminder).”
User 2: “No actions taken. Nothing is as frustrating to novel users as being forced
to enter data they are not interested in. It is even worse to prompt users fro
parameters they have no clue about. Everything possible should be done to
read all possible metadata out of the files themselves and to make metadata input
as painless as possible (templates, and/or copying metadata among images is
useful), but nothing should be forced. ”
http://openmicroscopy.org
OME2 System Layout
Metadata
Database
OMEDS
Services
OMEIS
Image
Repository
Microscope
Metadata and
Analytics
Binary Image Data
“Pixels”
Data Migration with OME XML
Metadata
Digital
Image
Acquisition
System
Metadata
Raw
Image
Raw
Image
Analytic
Results
OME XML File
My OME
(or other)
Your OME
(or other)
Goldberg et al (2005) Genome Biol. 6:R47
http://openmicroscopy.org
Using Data Hierarchies in OME
Jane’s Thesis
Images currently
being analyzed
Image
1
Image
2
Cell 1
Image
5
Projects
Images to
send to Ian
Image
4
Cell 2
Chromosome 1
Volume
Chromosome 12
Volume
Image
8
Images collected
Tuesday
Image
10
Cell 1
Image
12
Image
13
Cell 2
Chromosome 9
Volume
Chromosome 13
Volume
Datasets
Image
15
Images
Features
OME 2.x-- Java UI
OME 2.x-- Java UI
OME 2.x-- Web Browser UI
http://openmicroscopy.org
OME 2.x-- Web Browser UI
http://openmicroscopy.org
OME Technology
OME: Dirty Laundry
•Client-server transfer too slow for large
images & data graphs
•Import of external analysis not yet available
•Full image analysis suite not yet available
•UI’s are sometimes clunky
•Data deletion highly controlled
•Developers & users from Singapore to California
•Project runs naked
http://openmicroscopy.org
OME: Roadmap 2006 (Q1 - Q2)
•OMERO Release
•Release of improved OMEIS
•Improved display of complex data
•Integration of OME Data Model with MODs
•Flexible analysis capture
•5-dimensional annotation
http://openmicroscopy.org
The Boss
Jason Swedlow
The Lab
Ana Amores
Markus Posch
Paul Andrews
Elena Knatko
Iain Porter
Guennadi Khoudoli
Will Moore
Arwen Wilcock
Andrea Falconi
Jean-Marie Burel
Chris Allan
Joshua Moore
OME
Peter Sorger
Ilya Goldberg
Kevin Elicieri
John White
GRE
Angus Lamond
Julian Blow
Tomo Tanaka
Mike Stark
Tom Owen-Hughes
Neil Perkins
Joost Zomerdijk
Janice Walker
Univ of Southern Denmark
Jens Andersen
Matthias Mann
Univ of Pennsylvania
John Murray
CMB
Kate Storey
OME Commercial Partners
Bitplane
Applied Precision
PerkinElmer
Chroma Technology
Omega Optical
MD/UIC
Zeiss Confocal
OPIA
http://openmicroscopy.org