Reference Genome Project

1
Reference Genome Project
ZFIN
2
(No Transcript)
3
Purpose

• - Provide comprehensive annotation for 12
  genomes
• Arabidopsis thaliana
• Caenorhabditis elegans
• Danio rerio
• Dictyostelium discoideum
• Drosophila melanogaster
• Escherichia coli
• Gallus gallus
• Homo sapiens
• Mus musculus
• Rattus norvegicus
• Saccharomyces cerevisiae
• Schizosaccharomyces pombe

• Those organisms were selected because they are
• established model organisms with published
  experimental data
• have a genome database
• have experienced GO curators

4
Complete genome annotation

• Breadth every gene in the 12 genomes be
  annotated
• Depth every gene be annotated to the highest
  level of knowledge

• The group has agreed that depth of annotation
  is best assessed by the curator annotating the
  gene. - If a gene has less than 5-10 papers, it
  makes sense to read and annotate all papers
• - If a gene has a lot of literature, the
  preferred strategy is to look at a recent review
  to make sure all important primary literature is
  captured more recent papers be read

5
Metrics assessing breadth and depth of
annotations
Need GFF3 files for all Ref Genomes!

• Breadth
• Number of genes (protein coding and functional
  RNAs based on SO)
• Number of genes with some functional annotation
• Number of genes with functional annotation based
  on experiments using that organism
• Number of genes with function inferred by
  sequence similarity
• Number of genes with function inferred by
  electronic annotations
• Number of genes for which there is no available
  information (root/ND annotations)
• Depth
• Number of papers linked to a gene
• Number of papers used to produce functional
  annotation
• Number of papers read but for which no new
  annotations were produced.
• Ratio of deepest annotation to leaf node to
  measure granularity and use of the ontology (Suzi)

Depth needs to be assessed by curators!!
http//gocwiki.geneontology.org/index.php/Metrics
_breath_and_depth_of_annotations
6
Figures for Reference Genome genes, completely
fictitious data
Figures for Whole Genome genes, completely
fictitious data
From Ruth (word doc sent earlier this week)
7
Mike Cherry
8
(No Transcript)
9
Measuring Information Content
Organism Distance to leaf Distance to leaf Information Content Information Content Coverage Coverage Pubs per gene Pubs per gene Terms per gene Terms per gene
All Ref All Ref All Ref All Ref All Ref
Arabidopsis thaliana 4.29 3.74 11.53 12.01 17.06 28.28 2.02 2.20 2.88 3.46
Caenorhabditis elegans 4.74 4.37 11.40 11.90 27.98 48.94 1.73 3.23 4.72 9.51
Danio rerio 4.35 4.01 11.01 11.66 19.34 33.56 1.65 2.36 6.12 8.10
Dictyostelium discoideum 4.40 3.46 11.37 12.77 25.84 41.56 1.89 2.46 4.44 6.41
Drosophila melanogaster 3.56 3.03 12.86 13.37 33.29 55.69 3.12 5.34 4.26 6.55
H. sapiens 3.65 3.65 12.96 12.87 30.80 73.71 6.01 6.01 4.65 6.82
Mus musculus 4.83 3.50 10.90 13.37 32.44 84.21 2.95 10.69 5.47 13.01
R. norvegicus 4.08 3.26 11.56 13.10 36.27 77.81 2.29 5.71 5.07 8.00
S. pombe 3.82 3.05 11.94 12.96 44.15 59.54 3.13 4.01 6.91 17.55
Saccharomyces cerevisiae 3.22 2.63 12.88 14.08 37.42 62.84 3.04 5.67 4.65 6.82
Chris Mungall
10
Priorities Selection of curation targets

• Genes that, when mutated, cause a disease
• Not included upregulated in cancer x,
  interacts with tumor suppressor y, and other
  weak evidence
• Disease gene lists - OMIM- RGD disease portal
  first group neurological diseases
• Other lists - list of common genes between
  human, fly and zebrafish that were being used as
  a test case for PATO annotations many were not
  in OMIM revisit??
• Current status trying to focus on genes with
  the broadest interest, however these often lack
  orthologs in yeast, E. coli, etc, so need to
  balance these factors.

11
Orthologs

• Curators for each database are responsible for
  identifying orthologs of the disease gene
• Available tools - YOGY- InParanoid-
  OrthoMCL- TreeFam- Homologene
• Sequence analysis by curators

12
Software

• Google spreadsheet - shared by all curators-
  each database keeps track of putative orthologs-
  each database records the curation status for
  each gene
• Software requirements - Ensures consistent use
  of identifiers- Allow loading of MOD reports-
  Track that no ortholog was found- Provide
  reports to focus curation effort- Record that
  curation is 'comprehensive' as of a certain
  date- Allow a 1many relation between Human gene
  and MOD ortholog- Record orthology determination
  method

13
http//rails-dev.bioinformatics.northwestern.edu2
4000/index.html
14
Annotation Progress
Organism Genes with Ortholog Gene Curated Curated genes with publications
A. thaliana 32 99 32
C. elegans 65 46 99
D. discoideum 40 41 26
D. melanogaster 48 50 67
D. rerio 87 90 26
H. sapiens 44 98
M. musculus 98 84 94
R. norvegicus 96 100 55
S. cerevisiae 30 100 99
S. pombe 34 81 33

• Curation software will be able to generate that
  information
• We would like to display the list of selected
  genes, the list of identified orthologs, the
  curation status and a way to access annotations
  (graphs)

15
Annotation Consistency Comparing annotations
16
(No Transcript)
17
Ontology development
Number of Source Forge requests in the "Reference
Genome" group
18
Outreach publicizing the reference genome effort

• Several suggestions
• GO newsletter (already have the gene of the
  quarter) could add diseases
• NCBI/OMIM could display/advertise genes with
  annotations
• Take advantage of user requests that fit nicely
  in the initiative
• Set up a reference genome wiki page showing which
  genes are coming up for annotation, which could
  also be used by researchers to suggest target
  genes
• Make a page on the GO website that would include
  diseases genes we are curating and the gene of
  the quarter articles
• Special display in AmiGO
• Provide annotations in a separate file
• Mark disease genes specifically in MODs

http//gocwiki.geneontology.org/index.php/Outreach
_publicizing_the_project_and_developing_a_web_pre
sence