Growing the Tree of Life at SDSC: Kick-Off Meeting

11 1

Growing the Tree of Life at SDSC:Kick-Off Meeting

Mark A. Miller

San Diego Supercomputer Center

OTHER SDSC/MULTI-SITE PROJECTS in BIOLOGY

• Alliance for Cell Signaling

• Encyclopedia of Life

• Joint Center for Structural Genomics

• Protein Data Bank

• Protein Kinase Resource

• Lipid Maps (Lipid Metabolomics)

Informatics Center

VirtualCollectionsVirtual

CollectionsVirtual

Collections

Integrated Views

Integrated Views

Linux ClustersLocal

ResourcesLocal

Resources

Scientific Community

SMPs

Discovery Portal

Computational Resources

Domain ScientistsDomain Scientists

Disciplinary Interfaces• Molecular Biology• Structural Biology• Algorithm developers

Disciplinary Interfaces• Molecular Biology• Structural Biology• Algorithm developers

FederatedData

Collections

FederatedData

Collections

View Providers

View Providers

Tool Providers

Tool Providers

“The GRID”“The GRID”

TOL Project functions:

• Algorithm development• Model development• Methodologies for model evaluation• Database methodologies• Software development• Scalability/High end computing• Integration technology• Education/outreach

Role of SDSC:

• Unite the users with resources to do their science.• Create standards for the resource in partnership w/community.• Design a reference architecture that supports interoperability, evolvability, and extensibility.• Provide a common point of contact for project members and the community.• Create a resource that gives persistence to the work created under this funding.

For each project function we must ask:

• What are the overall goals?

• What resources do we have presently?

• Where is IT support limited?

• What are the requirements for delivering on goals?

• How can we formalize communication?

Of the project management we must ask:

How shall we prioritize the allocation of resources to goals?• How much effort to stabilize/wrap legacy SW (and which SW)?• What are trade offs of architectural specifications (SW and HW)?• What are the computational requirements (grid vs local) ?• What functions do we need to provide immediately?• What is the right mix of short term/long term investment?

Synergies: Features under development in other SDSC Projects

• HEC (DataStar: IBM power4; 7 tf/ 235 GB memory; SunFire 12K)

• Web/grid services (Blue Titan)

• Workflows (Ptolemy II; Scitegic)

• e-Notebook (partnership with Kinematik; internal efforts)

• Molecular visualization tools (mbt.sdsc.edu)

• Biological data federation (IBM DL federation;

http://quatermass.sdsc.edu:8080/federation/index.html )

Where do we start?

• Create infrastructure that supports communication.http://landscape.sdsc.edu:8080/PHYLO/

• Create a specifications document/project plan for each

functional area.

• Map the specifications to existing resources; strategize

around weaknesses.

• Propose one or more architectures that map to the groups

needs.

Proposed timelineNovember 2003 Inception:

• possible architectures described. • requirements for systems identified• project planning and staffing, • Life Cycle Objectives documentation with at least one feasible system architecture.

February 2004 Elaboration:• System Software Architecture Description created.• Life Cycle Architecture review and commitment.

October 2004 Construction• Major coding of product occurs, • culminating in the delivery of "alpha" capability to the user together with documentation.

SDSC Tree of Life Team

Mark Miller – Management/Logistics/Strategy

David Stockwell – Data Base/Architecture

Alex Borchers – Architecture/Data delivery tools/ User requirements

Dana Jermanis – Graphic Art/Web Master

1212 12

Metadata sharing

Virtual community messaging

TOL NotebookTOL Notebook

SOAP Queries

Database

SOAP Server

XML/RDF store

BLAST DataKeyword data

Stored queries

Annotations

Session info

Scheduler

BLASTKeyword queries

Invoke

Encyclopedia of Life