Baygenomics: The Bay Area Functional Genomics Consortium

1
BaygenomicsThe Bay Area Functional Genomics
Consortium

• Tom Ferrin
• and
• Patsy Babbitt
• University of California, San Francisco

2
Project Objectives

• 1. Use gene-trap vectors to inactivate 2,500
  genes/year in mouse embryonic stem (ES) cells
• As they are generated, all of these cell lines
  are being made freely available to the scientific
  research community
• 2. Identify those ES cell clones that have
  important roles in cardiopulmonary development
  and disease through in-situ hybridization and
  oligo-arrays
• 3. Produce knock-out mice from a selected subset
  of ES cells for further study and analysis

3
Organization

• Our efforts are built around 8 key components
• Gene Trapping in Embryonic Stem Cells
• Bill Skarnes, UCB
• Computational Methods for Predicting Gene
  Function
• Patsy Babbitt Tom Ferrin, UCSF
• In-situ Hybridization
• Pao-Tien Chuang, UCSF
• Gene-Expression Profiling and Analysis
• Bruce Conklin, Gladstone Foundation
• Mouse Resource for Lipid Metabolism and
  Atherogenesis
• Steve Young, Gladstone Foundation
• Mouse Resource for Pulmonary Disease
• Dean Sheppard, UCSF
• Mouse Resource for Cardiopulmonary Development
• Pao-Tien Chuang, UCSF
• Cardiopulmonary Genomics Education
• Patsy Babbitt Tom Ferrin, UCSF

4
Key Informatics Goals

• Provide, via the Baygenomics web site, access to
  all existing genetrap sequences
• Browse list of known genetrap genes
• Browse list of novel genes
• Download a single file of all sequence tags in
  FASTA format
• Search, via BLAST, user provided sequence(s)
  against genetrap sequences
• Order a cell line
• Develop and apply semi-automated bioinformatics
  protocols to store, annotate, and evaluate these
  genetrap sequences

5
(No Transcript)
6
(No Transcript)
7
Functional Analysis

• Functionally characterize gene-trap sequences of
  potential relevance through superfamily analysis
• Extend sequence fragments, when possible, with
  overlapping ESTs to form longer contigs
• Add functional definitions to sequences that have
  close homologues with known function
• Characterize ORFs whose only close homologues are
  of unknown function
• Characterize ORFs with only distantly related
  homologues
• Evaluate unknown ORFs for structural and
  functional characteristics using pattern- and
  motif-finding algorithms (e.g. Babbitts SHOTGUN
  program)

8
Status

• Programming staff hired (2 people at 50 time
  each) and up to speed
• Currently testing automated data upload
  capability of sequence and trace files generated
  at our genomic sequencing facility
• MySQL database populated with 400 initial
  genetrap sequences and by-hand annotations
• Initial database query scripts complete
• Dynamic HTML generation scripts still under
  construction

9
Status (contd)

• Staff Research Associate hired and working on
  automated annotation protocol
• Redesign of web site by graphics designer underway

10
Issues

• Functional analysis efforts would benefit from
  timely access to mouse genomic data
• Experimental components already up to speed, but
  software development components have only begun
  recently (after personnel hired) and will take
  time to complete
• Difficult to get the experimentalists to
  appreciate the time required to develop needed
  software infrastructure

11
Future Directions

• Integrate tissue in-situ hybridization images
  into MySQL database
• Add links to gene-expression profiling data being
  generated/compiled by Bruce Conklin (component 4)