Transcript From Functional Genomics to Physiological Model: the

Prioritization of Avian
GO Annotation
Structural Annotation
Genome No. Entrez
Species Build2
Genes
Human
Mouse
Rat1
Chicken
36.3
37.1
3.4
2.1
36,437
64,018
49,516
19,9793
No. Proteins % predicted
(NRPD)
proteins
415,830
228,696
108,069
31,8193
4.91
9.28
29.99
46.624
proteins/gene
11.41
3.57
2.18
1.595
NRPD: Non-redundant Protein Database
1. The
rat genome was published only 8 months prior to the chicken genome, yet rat has
2x as many genes in Entrez Gene and 3x as many proteins.
2. After two genome builds chicken still has 5% of genomic sequence that has not been
assigned a chromosome and mini-chromosomes have not been sequenced.
3. Chicken genes and proteins are under-represented in public databases.
4. Of the chicken proteins available from NRPD, almost half are predicted based upon
computational analysis.
5. On average chicken has only 1 protein per gene so very little is known about isoforms
and alternate transcripts in the chicken gene products.
Phase 1: “Breadth”

7, 478 Chicken entries in UniProtKB
 GOA provides
IEA mapping for UniProtKB entries
Initial strategy for AgBase biocurators was to
add GO to chicken gene products that had
none.
 Since 46% of the chicken proteins in NRPD
were predicted, they would have no GO

 IEA,
ISS, ISO….
Functional Annotation
100
80
% of gene
products
annotated
no GO
60
computational GO
AgBase
40
manual GO
20
0
Human
Mouse
Rat
Chicken
the proportion of GO for chicken is over-represented
because of their under-representation in public databases
Phase 2: “Depth”
What are the
community needs?
GO Annotation of Arrays
DelMar14K, FHCRC, Tgu array
 44K Agilent oligo array
 AIIM array, Affymetrix

Should we be focusing on arrays?
 What arrays should we do?

GO Annotation Priorities?
Provide “breadth” of coverage
 Annotate products represented on arrays
 Reference Genome targets
 Subject areas (immunity,
nutrition/metabolism, development
 Ad hoc as requested
