A Virtual Laboratory for Global Biodiversity Analysis

1
A Virtual Laboratory for Global Biodiversity
Analysis
2
Talk Outline

• Background Project Information
• Biodiversity World System
• Current Progress and Future Work
• Q (hopefully) A

3
Project Participants

• Southampton
• Oliver Bromley
• Cardiff
• Alec Gray
• Andrew Jones
• Richard White
• Nick Fiddian
• Xuebiao Xu
• Nick Pittas
• Bristol
• Paul Valdes

• Reading
• Frank Bisby
• Alistair Culham
• Neil Caithness
• Tim Sutton
• Peter Brewer
• Chris Yesson
• NHM
• Malcolm Scoble
• Paul Williams
• Shonil Bhagwat

4
Project Aims

• To create a prototype problem-solving
  environment (PSE) for Global Biodiversity
  research on the GRID.
• To demonstrate application of the prototype in a
  range of data/computation intensive biodiversity
  investigations.

5
Objectives

• to establish a biodiversity GRID with nodes at
  Reading, Cardiff, Southampton and the NHM
• to design the architecture of a GRID-based PSE
• to build, and test the basic system
• to demonstrate the system in use for three
  exemplar analyses.

6
Computer Science Challenges

• To achieve the seamless integration of the
  resources required in order to construct the PSE
• Deal with heterogeneity of resources
• Accommodate for the complex analyses required
• Consider metadata formats for selecting and
  interpreting data from appropriate resources
• Apply the above in a GRID environment

7
Exemplars

• BioClimatic Modelling Climate Change
• BioDiversity Richness Conservation Evaluation
• Phylogenetic Analysis Biogeography

8
BioClimatic Modelling

• Predicting species distributions under past,
  present and future climate scenarios.
• Models
• GARP (Genetic Algorithms for Rule-set Production)
• CSM (Climate Space Models)
• Bioclim

9
(No Transcript)
10
BioClimatic Modelling (cont)

• Has the plant already reached all suitable
  environments world-wide, or are further
  expansions possible?
• In which parts of Australia might it be worth
  introducing the plant?
• Where will this plant die out, and where might it
  now appear, if the world is subject to some of
  the global warming scenarios?

11
Biodiversity Richness Conservation Evaluation

• Concerned with analysis of biodiversity richness
  patterns for particular taxa around the globe.
• Different approaches to measuring biodiversity
  (by species richness or by taxic diversity)
  depending on the purposes for which the measures
  are required.
• WORLDMAP to be used as the analysis software (NHM)

12
(No Transcript)
13
Biodiversity Richness Conservation Evaluation
(cont)

• Enhance conservation network design by answering
  questions about patterns of complementarity
  (species difference among areas).
• Provide biodiversity richness assessment for the
  Geometer Moths group.

14
Phylogenetic Analysis Biogeography

• Aims to use phylogeny to interpret biodiversity
  data such as
• Species distribution,
• Species morphology, and
• Life history evolution.

15
Phylogenetic Analysis Biogeography (cont)

• is geography a good predictor of relationship
  among lineages?
• do all lineages show the same dispersal capacity?
  
• have lineages stayed put, adapting in situ while
  climates have changed?

16
Architecture (simplified)
17
Data Flow
18
Resource Modules

• Two types of resources
• Analytic resources (services)
• Data resources.
• Wrapped through standard communication Interface.
• Thus BDWorld GRID split into 2 sub-Grids
• Computational GRID
• Data Grid

19
Workflow Manager

• Main point of entry for the user into the system.
• Allows the user to define the sequence of tasks
  (with associated data) in order to complete an
  analysis.
• Two versions investigated in parallel
• Current one based on XPDL representation and the
  Open Business Engine (OBE) WF engine.
• Has been decided to revert to the TRIANA workflow
  engine (if not both the TRIANA WFM engine and UI)
  in the near future.

20
Metadata Repository

• Allows for resources to publish their metadata
• Computational capabilities
• Supported data types
• Allow for workflow (sub)sequences to verify their
  validity
• Holds advanced system information
• Data provenance
• Alternative computational resources for a given
  task
• Currently implemented as a relational DB
• Plans to move on a semantically more flexible
  implementation (OAVs) in future releases

21
Ontology

• Provides a high level description of system
  entities
• Helps the user in workflow formulation
• Concept hierarchies that denote equivalent
  concepts/resources
• Concept association showing intra-concept
  relationships
• Currently work in progress

22
Communications Layer API

• Allows for communication among system components
• Remote component invocation
• Data interchange
• Provides clients that invoke resources
  transparently
• Provides server side behaviour to resources by
  inheriting a single class

• Implements a standard Data Exchange format
  (Object ?XML serialisation/revival).
• Allows for monitoring the progress of active
  processes.
• Currently existing in two flavours
• RMI-based one (pre-GLOBUS era)
• OGSA based.

23
Communications Layer API (cont)
24
Using Globus

• No major problems encountered (post-beta
  versions)
• Tricky to work with complex data structures
• Deployment of finished resources is
  time-consuming/error-prone.

25
Current Progress

• RMI communications layer currently used
• OGSA based architecture about to be rolled out
  (was?)
• Enough resources to carry out the Bioclimatic
  analysis exemplar. (Biogeography almost ready)
• Rudimentary Metadata repository exists.
• XDPL/OBE Workflow manager has been successfully
  implemented. TRIANA currently considered.

26
Future plans

• Introduce systems Ontology
• Revert to TRIANA
• Convert to an all OGSA architecture(??)
• Introduce resources/workflows for the remaining 2
  exemplars.
• Enhance the representational power of the
  metadata repository.

27
Questions ?