GENEBOREE

1

• GENEBOREE
• A Tool for Collaborative Gene Annotation
• DOE Joint Genome Institute
• Integrated Microbial Genomes
• Annette Greiner
• Anya Kartavenko
• Rowena Luk

2
A Genetics Primer
3
Geneboree

• Collaborative gene annotation
• Geographically distributed
• Open to all bioscientists

4
Goals

• Credibility
• Security
• Standards
• Ease of use
• Seamless integration

5
Developing Personas

• Interviews at JGI
• Professional annotators (3)
• Scientists (1)
• Interviews with field scientists
• UC Berkeley (2)
• LBL (3)

6
Personas

• Bill Newman, PhD, 48
• Staff Scientist at Los Alamos National Lab
• Microbial bioterrorism
• Naïve computer user and recognized authority in
  his field

• Martha Jane Gilbert, PhD, 30
• Postdoctoral fellow at University of Oxford, UK
• Metagenomics
• Intermediate computer user and fairly experienced
  professional

• Phil Dupont, 24 (Primary User)
• Third year PhD student at Rutgers University
• Fungi studies in evolutionary biology
• Expert computer user and least experienced
  professional

7
Fourth Persona

• Bill Newman, PhD, 48
• Staff Scientist at Los Alamos National Lab
• Microbial bioterrorism
• Naïve computer user and recognized authority in
  his field

• Martha Jane Gilbert, PhD, 30
• Postdoctoral fellow at University of Oxford, UK
• Metagenomics
• Experienced computer user and fairly experienced
  professional

• Phil Dupont, 24 (Primary User)
• Third year PhD student at Rutgers University
• Fungi studies in evolutionary biology
• Expert computer user and least experienced
  professional

• Ann Taylor, 35
• Professional annotator at Sanger Institute, UK
• Updates automatic annotations
• Intermediate computer user and expert annotator

8
Scenario Bill

• Compares gene sequences from Chernobyl and finds
  an oddball
• Verifies annotators credibility
• Updates annotation

9
Scenario Martha Jane

• Searches IMG for genes involved in carbon
  fixation
• Notices that most results were automatically
  annotated and not reliable
• Makes a comment on all of the genes in her set

10
Scenario Phil

• Decides to transfer annotation from one homolog
  to his gene of interest
• Just to be sure, he checks his assumptions
  against a phylogenetic tree and finds that the
  annotation he wanted to transfer was inaccurate
• Corrects annotation of the homolog and annotates
  his gene of interest

11
First Interactive Prototype

• Built with Dreamweaver/PHP/MySQL
• Dreamweaver templates control navigation, headers
• PHP and MySQL make pages dynamic

12
First Interactive Prototype

• Prototype uses the existing IMG site for existing
  IMG functions
• IMG (http//img.jgi.doe.gov)
• Gene searches
• Comparative tools
• Geneboree (http//groups.sims.berkeley.edu/annotat
  ion)
• Collaborative annotation
• Discussion about gene function
• Authentication

13
First Interactive Prototype

• Differences from paper prototype
• No drafts
• Replies to batch discussion
• Automatic previews (can be turned off in prefs)
• Voting buttons on gene cart page
• Hyperlinks in discussions
• Flags to indicate genes recently annotated
• New method of comparing and transferring
  annotations from other genes
• Additional evidence fields (Pfam and UniProt)

14
First Interactive Prototype
15
First Interactive Prototype

• http//groups.sims.berkeley.edu/annotation/

16
Lessons Learned

• Users are different!
• Manual checking vs automatic annotation
• Homology vs. phylogeny vs. experimental data
• More evidence
• Open question does voting work?

17

• Questions?