My Collaborators

1
My Collaborators Other Animals

• Alastair Knowles
• (Incognizant) Project Manager
• alastair.knowles_at_ncl.ac.uk

2
What is CARMEN?

• e-Science Pilot Project (4.5M October 2006)
• 11 UK Universities to develop a virtual
  laboratory for neurophysiology
• Development of core infrastructure at Newcastle
  and York
• Partners from Plymouth to St. Andrews conducting
  neurophysiology

3
Neurophysiology Data

• The study of nervous system function
• Primarily electrode recordings
• Some imaging using magnetic and labelling
  techniques
• Experiments in human and animal tissue
• GBs/hour

10-25nm
4
Neurophysiology Analysis

• 64/128/256/512 channels, 25,000 samples/s
• Spike detection and sorting
• Identification of higher order correlations and
  patterns
• Information theory
• Network analysis
• Models

Signal
Spikes
5
CARMEN System Architecture
6
CARMEN Team Architecture
Developers Software
Developers Metadata
Project Manager Communication
End User Experiments
End User Models
End User Analysis
7
Timescales

• Release of data sharing infrastructure (imminent)
• Release of analysis platform (Q1/Q2 2009)

8
Issues

• Sharing
• Metadata and annotation
• Ownership and attribution
• Community
• Sustainability

9
Sharing

• Like courtship
• Small secretive groups with wide ranging
  territory
• Difficult to find partners
• Once found, dont let somebody else find them!
• Some partners were reluctant to
  describe/disclose collaborations with individuals
  external to the consortium. Further, a minority
  of partners were protective of their
  collaborations, e.g. reluctant to introduce
  collaborators including CARMEN partners to other
  prospective collaborators within the consortium.
  Alastair Knowles Project Review Sept 2008.

10
Sharing

• Are there any gold standard datasets that would
  encourage you to collaborate in CARMEN?
• Partners were asked to specify third party
  resources (e.g. resources external to the
  project) that they would like to be made
  available to them in CARMEN. The majority
  identified resources belonging to their immediate
  collaborators external to CARMEN. A minority
  identified their own research as providing the
  most critical resources. There are no gold
  standard resources. Alastair Knowles Project
  Review September 2008.

Collaborators
Mine!
11
Metadata and Annotation Genome

• Triple helix (1953)
• DNA Sequencing (1976-80)
• Los Alamos sequence database (1979)
• GenBank (1982)

12
Metadata and Annotation - Neurophysiology

• Need to exchange lots of rich, freeform
  information
• Cannot be constrained by structure
• Need to find the needle in the haystack
• Need to maintain individuality
• Lingua franca MINI (minimum information about a
  neurophysiology investigation)
• http//www.carmen.ncl.ac.uk/standards/mini.pdf

13
Ownership and Trust

• Trust
• Protracted ownership
• Just make everything open?
• What species are you?

Analysis
Data 2
Data 1
Data 3
14
Community

• Big picture?
• Journals, funders?

15
Sustainability

• Data sharing policy
• Support for long term users
• High cost solutions for niche audiences?

16
In the Ideal World ...
Science Cloud
17
CARMEN Consortium