1. Introduction to the Gene Ontology and GO annotation resources
Rachael Huntley
UniProt-GOA
EBI
Programmatic Access to Biological Databases
3rd October 2012

2. What is an Ontology?
What is the difference between an ontology and a controlled vocabulary?

3. An ontology formally represents knowledge as a set of concepts within a domain, and the relationships among those concepts. It can be used to reason about the entities within that domain and may be used to describe the domain.
A Controlled vocabularies provide a way to organize knowledge for subsequent retrieval but does not allow reasoning about the entities.

4. Ontologies and CVs in Biology
Ontologies/CVs are heavily used in biological databases
Allow organisation of data within a database
Enable linking between databases
Enable searches across databases

7. What is GO?

8. The Gene Ontology A way to capture
Less specific concepts
A way to capture
biological knowledge for individual gene products
in a written and
computable form
A set of concepts
and their relationships
to each other arranged
as a hierarchy
More specific concepts

9. The Concepts in GO 1. Molecular Function 2. Biological Process
protein kinase activity
insulin receptor activity
An elemental activity or task or job
2. Biological Process
A commonly recognised series of events
cell division
mitochondrion
mitochondrial matrix
mitochondrial inner membrane
3. Cellular Component
Where a gene product is located

10. Anatomy of a GO term Unique identifier Term name Synonyms Definition
Cross-references

11. Ontology structure Directed acyclic graph Terms are linked by
Terms can have more than one parent
Terms are linked by
relationships
is_a
part_of
regulates (and +/- regulates)
has_part
occurs_in
These relationships allow for complex analysis of large datasets

12. Relations Between GO Terms
is_a
If A is a B, then A is a subtype of B
part_of
Wherever B exists, it is as part of A. But not all A is part of B.
All replication forks are part of a chromosome
Not all chromosomes have replication forks A B

13. Relations Between GO Terms Regulates
One process directly affects another process or quality
Necessarily regulates: if both A and B are present, B always regulates A, but A may not always be regulated by B A B
All cell cycle checkpoints regulate the cell cycle.
The cell cycle is not solely regulated by cell cycle checkpoints

14. Process-Function Links in GO
GO was originally three completely independent hierarchies, with no relationships between them
Biological processes are ordered assemblies of molecular functions
As of 2009 we have started making relationships between biological process and molecular function in the live ontology
functions that regulate processes e.g. transcription regulator regulates transcription
process
process
function
functions that are part_of processes e.g. transporter part_of transport
function

15. Searching for GO terms Search GO terms or proteins

16. Exercise
Search for a GO term Exercise 1 (pg.16)

17. Why do we need GO?

18. Reasons for the Gene Ontology
Inconsistency in English language

19. Inconsistency in English languauge
Same name for different concepts
Cell or
The English language is not precise ??

20. Eggplant Brinjal Aubergine Melongene Same for biological concepts
Different names for the same concept
Eggplant
Brinjal
Aubergine
Melongene
Same for biological concepts
Comparison is difficult – in particular across species or across databases
Just one reason why the Gene Ontology (GO) is is needed…

21. Reasons for the Gene Ontology
Inconsistency in English language
Increasing amounts of biological data available
Increasing amounts of biological data to come

22. Increasing amounts of biological data available
Search on ‘DNA repair’...
get over 68,000 results
Expansion of sequence
information

23. Reasons for the Gene Ontology
Inconsistency in English language
Increasing amounts of biological data available
Increasing amounts of biological data to come
Large datasets need to be interpreted quickly

24. • Compile the ontologies
Aims of the GO project
• Compile the ontologies
- currently over 38,000 terms
- constantly increasing and improving
• Annotate gene products using ontology terms
- around 30 groups provide annotations
• Provide a public resource of data and tools
- regular releases of annotations
- tools for browsing/querying annotations and editing the ontology

25.
Reactome

26. GO Annotation

27. UniProt-Gene Ontology Annotation (UniProt-GOA) project at the EBI
Largest open-source contributor of annotations to GO
Provides annotation for more than 350,000 species
Our priority is to annotate the human proteome

28. A GO annotation is … …a statement that a gene product;
has a particular molecular function
or is involved in a particular biological process
or is located within a certain cellular component
as determined by a particular method
as described in a particular reference
P00505
Accession
Name
GO ID
GO term name
Reference
Evidence code
IDA
PMID:
aspartate transaminase activity
GO:
GOT2

29. UniProt-GOA incorporates annotations made using two methods
Electronic Annotation
Quick way of producing large numbers of annotations
Annotations use less-specific GO terms
Only source of annotation for many non-model organism species
Manual Annotation
Time-consuming process producing lower numbers
of annotations
Annotations tend to use very specific GO terms

30. Electronic annotation methods
Mapping of external concepts to GO terms
GO: : Nucleus
GO: : MAP kinase activity
GO: : Auxin mediated signaling pathway

31. Electronic annotation methods
2. Automatic transfer of manual annotations to orthologs
Ensembl compara
Macaque
Chimpanzee
Guinea Pig
Rat
...and more
e.g. Human
Mouse
Cow
Dog
Chicken
Arabidopsis
Ensembl compara
…and more
Electronic annotations are important for non-model species as these do not have a dedicated manual annotation effort
Poplar
Brachypodium
Maize
Grape
Rice
Annotations are high-quality and have an explanation of the method (GO_REF)

32. Manual annotation by GOA
High–quality, specific annotations made using:
Full text peer-reviewed papers
A range of evidence codes to categorise the types of evidence found in a paper
e.g. IDA, IMP, IPI

33. Number of annotations in UniProt-GOA database
Electronic annotations
102,205,043
1,149,802
Manual annotations*
Sep 2012 Statistics
* Includes manual annotations integrated from external model organism and specialist groups

34. How to access and use
GO annotation data

35. Where can you find annotations?
UniProtKB
Ensembl
Entrez gene

36. UniProt vs. QuickGO annotation display

37. UniProt vs. QuickGO annotation display
Filtering mechanism
Exclude root terms (e.g. Molecular Function)
Exclude annotations with qualifier (e.g. NOT, contributes_to)
Exclude annotations to less granular terms
Exclude annotations to GO: protein binding
Exclude lower quality assignments for same data, e.g. UniProt taken in preference to MGI
Add electronic annotations that cover ground not covered by manual annotation

38. Gene Association Files
17 column files containing all information for each annotation
GO Consortium website
UniProt-GOA website
Numerous species-specific files

39. GO browsers

40. The EBI's QuickGO browser
Search GO terms
or proteins
Find sets of
GO annotations

41. Exercise Find annotations to a protein Exercise 2 (pg.16)
Find annotations to a list of proteins Exercise 1 and 2 (pg.22)

42. How scientists use the GO
Access gene product functional information
Analyse high-throughput genomic or proteomic datasets
Validation of experimental techniques
Get a broad overview of a proteome
Obtain functional information for novel gene products 
Some examples…

43. Term enrichment Most popular type of GO analysis
Determines which GO terms are more often associated with a specified list of genes/proteins compared with a control list or rest of genome
Many tools available to do this analysis
User must decide which is best for their analysis

44. Analysis of high-throughput genomic datasets
attacked
time
control
Puparial adhesion
Molting cycle
Hemocyanin
Defense response
Immune response
Response to stimulus
Toll regulated genes
JAK-STAT regulated genes
Amino acid catabolism
Lipid metobolism
Peptidase activity
Protein catabolism
MicroArray data analysis
Genes differentially expressed during parasitoid attack versus control
Bregje Wertheim at the Centre for Evolutionary Genomics, Department of Biology, UCL and Eugene Schuster Group, EBI.

45. Annotating novel sequences
Can use BLAST queries to find similar sequences with
GO annotation which can be transferred to the new sequence
Two tools currently available;
AmiGO BLAST – searches the GO Consortium database
BLAST2GO – searches the NCBI database

46. Using the GO to provide a functional overview for a large dataset
Many GO analysis tools use GO slims to give a broad
overview of the dataset
GO slims are cut-down versions of the GO and
contain a subset of the terms in the whole GO
GO slims usually contain less-specialised GO terms

47. Slimming the GO using the ‘true path rule’
Many gene products are associated with a large number of descriptive, leaf GO nodes:

48. Slimming the GO using the ‘true path rule’
…however annotations can be mapped up to a smaller set of parent GO terms:

49. GO slims Custom slims are available for download;
or you can make your own using;
QuickGO
AmiGO's GO slimmer

50. The EBI's QuickGO browser
Search GO terms
or proteins
Find sets of
GO annotations
Map-up annotations
with GO slims

51. Precautions when using GO annotations for analysis
The Gene Ontology is always changing and GO annotations are
continually being created
- always use a current version of both
- if publishing your analyses please report the versions/dates you used
Recommended that ‘NOT’ annotations are removed before analysis
- only ~3000 out of 57 million annotations are ‘NOT’
- can confuse the analysis

52. Precautions when using GO annotations for analysis
Unannotated is not unknown
- where there is no evidence in the literature for a process, function or
location the gene product is annotated to the appropriate ontology’s
root node with an ‘ND’ evidence code (no biological data), thereby
distinguishing between unannotated and unknown
Pay attention to under-represented GO terms
- a strong under-representation of a pathway may mean that normal
functioning of that pathway is necessary for the given condition

53. Exercise Use the QuickGO web services to retrieve annotations
Exercise 1 Pg.46

54. The UniProt-GOA group Project leader: Rachael Huntley Curators:
Yasmin Alam-Faruque
Prudence Mutowo
Software developer:
Tony Sawford
Team leaders:
Rolf Apweiler
Claire O’Donovan

55. Acknowledgements Members of; UniProtKB InterPro Ensembl IntAct
HAMAP
Ensembl
Ensembl Genomes
GO Consortium
Funding
National Human Genome Research Institute (NHGRI)
British Heart Foundation
EMBL