CellML

1
CellML Synthetic Biology

• James Lawson
• Friday 16th May 2008

2
Overview

• Rundown on what the CellML team do here
• What is synthetic biology?
• What does CellML have to do with it?

3
ABI Team
We have grown!
(arrows show people who joined last year)
4
2001 - present
CellML is intended to support the definition of
models of cellular and subcellular processes.
CellML facilitates the re-use of models and parts
of models by using a component-based
architecture. Models are split into logical
sub-parts called components that are connected
together to form a model. CellML 1.0
specification
5
What is CellML?

• Language for description of mathematical models
• Developed at ABI (mainly)
• Open XML standard
• Describes structure, mathematics, metadata
• MathML for mathematics
• RDF for metadata

6
Why is there a need for CellML?

• Designed for storage exchange of models
• Reference implementation of a model
• Publication of model code
• Model reuse

ERRORS
A model is created
and translated into parameter sets and equations
then read and interpreted
and finally reimplemented
7
Why is there a need for CellML?

• If the model author published their model in
  CellML, this process could be avoided

ERRORS
A model is created
and translated into parameter sets and equations
then read and interpreted
and finally reimplemented
8
Scope of CellML

• CellML 1.1 current version, 1.2 in pipeline
• Flexible not limited to biological systems
• Philosophy
• Only describes structure and maths. All else is
  metadata.
• Multiscalar from biochemistry to physiology

9
Signal transduction
Electrophysiology
Biomechanics
CellML can describe multiscalar systems
10
IUPS Physiome Project
Nickerson et al. Computational multiscale
modeling in the physiome project modeling
cardiac electromechanics., IBM J. Res. Dev.,
50(6), 617-630, 2006
11
PCEnv

• Physiome CellML Environment
• Developed at ABI
• Open source / free environment for editing and
  simulating CellML models
• Uses CellML API
• Built using Mozilla/XUL framework
• Platform agnostic

12
(No Transcript)
13
CellML Model Repository

• www.cellml.org/models started life as set of
  test-cases for CellML
• Now a repository of gt330 unique CellML models
  based on peer reviewed publications
• Approximately half curated
• Models can be uploaded concurrent with
  publication
• Locus for sharing and reuse of models

14
(No Transcript)
15
(No Transcript)
16
Model Curation

• Important element of the work done in the CellML
  community
• Purpose is to provide validated, tested reference
  implementations of models
• Errors produced in code gt publication gt
  reproduction process must be eliminated
• Metadata and documentation
• Ontological annotation
• Authorship, citation and revision histories

17
CellML in CMISS

• Use CellML models to inform cell behaviour in
  heart (and other) models
• John Davidson, Jesse Ashton, Jichao Zhao
• Know more than me
• Are very friendly
• Are not the only people doing this

18
CellML Community

• Active international community of developers and
  users
• Cellml.org serves as focal point
• CellML specifications
• CellML Model Repository
• User profiles
• Hosts proposals, wikis, tutorials, FAQs etc.
• Meeting minutes, news, conference proceedings
• Software downloads
• Tracker

19
Projects in the Pipeline

• CellML 1.2
• Metadata
• Working with CellML in PCEnv
• The new CellML Model Repository PMR2
• Using CellML plans for community and curation

20
CellML 1.2

• Latest stable version of CellML is 1.1
• Last changed in 2002, frozen in 2006
• Community canvassing process
• Purification of CellML according to philosophy

21
CellML 1.2

• CellML too large to be implemented in entirety
• gt concept of secondary specifications which
  represent a subset of CellML
• Individual tools can implement these in entirety
• E.g. stochastic CellML

22
CellML 1.2

• Reaction element removal
• Ontologies and constrained vocabularies
  increasing CellMLs specificity in a modular
  fashion CellMLBio Ontology
• New typing system
• Implementation of a complex typing system
• Multiple subsets each subset introduces one core
  functionality
• E.g. vectors, matrices, operations over these
• Lambda calculus being considered

23
CellML Metadata

• Metadata specification to be frozen
• Development of CellMLBio Ontology
• Annotation of variables and components with
  biologically relevant information
• Automatic generation of diagrams of CellML models
• Use of graphing simulation metadata to
  reproduce figures from publications

24
(No Transcript)
25
PCEnv

• PCEnv development strongly coupled to development
  of CellML API
• Ability to link to and display references to web
  accessible databases
• Further support for metadata editing and
  processing
• Rendering of and user interaction with model
  diagrams
• Non-linear solver

26
CellML Model Repository PMR2

• Current version of CellML Model Repository does
  not adequately handle multi-file CellML 1.1
  models
• PMR2 currently in development
• Current prototype software uses distributed
  version control system, will allow full revision
  histories, treatment of CellML as code.
• Will allow models in the repository to reference
  each other using import element

27
PMR2

• Facilitates decomposition of models into modular
  components.
• Models are then composed of networks of these
  components and descriptions of the relationships
  between these.
• Curated, ontologically annotated components able
  to be searched, sorted, combined

28
Future cellml.org community

• Web-based communities are becoming commonplace
• International scientific collaboration demands
  effective communication media
• Project-centric workspaces
• Meeting minutes, project descriptions, seminars /
  posters, software downloads

29
Synthetic Biology
30
(No Transcript)
31
Synthetic Biology

• Wikipedia A new area of research that combines
  science and engineering in order to design and
  build ("synthesize") novel biological functions
  and systems
• SB1.0 held in 2004, SB4.0 in Hong Kong, October
  this year
• Synthetic biology tipped to be key technology of
  early 21st century

32
True Genetic Engineering

• An engineering approach to genetic engineering!
• Abstraction
• Standardisation
• Modularity
• CAD
• Refinement of natural products
• Resources need to be refined before their
  behaviour can be reliably described in
  calculations
• MIT has lead the charge with their Registry of
  Standard Biological Parts

33
(No Transcript)
34
(No Transcript)
35
You might also have heard the term Biobrick
What is a Biological Part?
Definition is shaky, but A nucleic acid
sequence that encodes a definable biological
function Shetty et al. 2008
Specification information for a Biological Part
looks like it was written by an engineer.
36

• International Genetically Engineered Machine
  competition
• Undergraduate competition started in 2004
• Central to current synthetic biology community
• 2008 53 teams registered, 31 pending registration

37
What has this got to do with us?

• I have been following synthetic biology since
  about mid 2005
• Mike Cooling is also very interested
• We have been sneakily bouncing ideas off each
  other in our spare time

38
CellML Synthetic Biology

• Synthetic biology currently DNA-centric, with
  much time spent cloning genes in the lab
• We propose that more can be done in silico
• We are interested in creating a Registry of
  Standard Biological Models
• We are collaborating with researchers from MIT,
  Imperial College and Newcastle to create a
  repository of modular, reusable curated models
  c.f. the CellML Model Repository

and write a paper about it
39
So Mike and I recently visited the UK
40
And we had a great time

• We visited Neil Wipats group at the School of
  Computing Science at Newcastle University
• We also attended the BioSysBio 2008 conference at
  Imperial College, London
• Focuses on systems and synthetic biology

You may recognise Neil he spent his sabbatical
at the ABI over the summer. During this time Mike
and I got to know him and started doing some work
together.
41
Newcastle University

• Mike and I gave a tag team talk
• Mike on modularity in CellML
• My talk was a general CellML now and in the
  future overview
• Newcastle is fielding a 2008 iGEM team
• Their team has a heavy focus on modeling of
  Standard Biological Parts
• CellML is one of the chief technologies involved
  in their project
• Mike and I are official advisors to their team

42
Newcastle University

• Bio-ontologies
• Philip Lord is a bio-ontologies expert
• Was very interested in the CellML-Bio Ontology
• Much important feedback received regarding CellML
  associated services
• E.g. Java bridges to API, programmatic access to
  CellML repository, stochastic modeling in CellML

43
(No Transcript)
44
(No Transcript)
45
BioSysBio2008

• Academic content of keynotes was average
• Workshops and quality of attendees 10/10
• Interesting to see what people knew / thought of
  CellML
• Physiome Project, CellML, Peter Hunter got big
  mention in introductory talk of conference
• Many people knew little about CellML and simply
  saw it as SBMLs poor cousin

46
Standards in Synthetic Biology

• Dedicated workshop at BioSysBio
• Mike and Neil presented on why Standard Parts
  need to be modeled and why CellML is the best
  technology for this
• Explicit modularity very important
• SBML is incumbent but lacking
• Synthetic Biology ontology being developed

47
Open Science

• Open access journals BMC
• Open notebook science
• Universal Scientific Brownie Points
• Problem of rewarding non-publication
  contributions such as blog posts, reviews etc.
• Movement may have ramifications for ABI
• Watch this space

48
Synthetic Biology CellML

• Mike and I are working on a paper with Neil
  Wipat, Barry Canton from MIT and Vincent Rouilly
  from Imperial College
• Aim to model Standard Biological Parts in a
  manner that allows these models to be combined
  and simulated
• Aim to provide a curated repository of these
  models

49
Synthetic Biology ABI

• Synthetic biology is primarily an engineering
  discipline. Currently wet-lab oriented but
  starting to involve more in silico work
• Will have vast industrial relevance within 10
  years
• ABI needs to consider whether it wants to get
  involved
• iGEM team could be perfect 4th year project
• Upcoming ABI wet lab capabilities could be used
• Collaboration with wet-lab specialists required

50
Acknowledgements

• Peter Hunter Poul Nielsen for allowing me to
  indulge my imagination
• The CellML team
• Mike Cooling