Taxon and GO

1
Taxon and GO
2
Removal of sensu

• Old system
• larval development (sensu Insecta)
• def The process whose specific outcome is the
  progression of the larva over time, from its
  formation to the mature structure. As in, but not
  restricted to Insecta.

3

• New system
• instar larval development
• def The process whose specific outcome is the
  progression of the larva over time, from its
  formation to the mature structure. This begins
  with the newly hatched first-instar larva,
  through its maturation to the end of the last
  larval stage

4
GO is species-neutral

• The GO can support annotation across all species.
  
• This is good!

• Total exclusion of taxon information can
• create problems.

5

• We have 23802 terms.
• Time-sink for users.
• Worst for
• peripheral
• species groups.

http//bestblog.mlblogs.com/inside_pitch/images/ne
edle_in_the_haystack_4.jpg
6
Typical User Query Hello, I'd like to load
into our database only GO terms that are
relevant for plants. Is there such GO-slim ? If
not, is there a way to filter out GO terms that
are not applicable for plants? Thanks, -Naama
Menda
Plant and Animal Genome Meeting 2005 Speaker
Some people think the size of GO
makes it hard to handle. Audience Raucous
laughter and shouts of no really?
Sanger microbial annotation We need a
prokaryote slim or we will not be able to
annotate fast enough to meet our targets.
(paraphrased)
7
GO language can be quite subtle.

• membrane of vacuole
• with cell cycle-independent morphology
• def The lipid bilayer surrounding a vacuole that
  retains the same shape regardless of cell cycle
  phase. The membrane separates its contents from
  the cytoplasm of the cell.

8
To an outsider this is like

• (8) If the Secretary of State considers it
  necessary or desirable to do so, he may by
  regulations prescribe a scale or scales of fees
  to have effect, for such period as is specified
  in the regulations, in place of any scale or
  scales prescribed by the Commission and, if he
  does so, references in subsections (3) and (4) to
  the appropriate scale are to be read as respects
  that period as references to the appropriate
  scale prescribed by the Secretary of State.

Audit Commission Act 1998 - 1998 CHAPTER 18 -
http//www.opsi.gov.uk/acts/acts1998/ukpga_1998001
8_en_2
9
To an outsider this is like

• (8) If the Secretary of State considers it
  necessary or desirable to do so, he may by
  regulations prescribe a scale or scales of fees
  to have effect, for such period as is specified
  in the regulations, in place of any scale or
  scales prescribed by the Commission and, if he
  does so, references in subsections (3) and (4) to
  the appropriate scale are to be read as respects
  that period as references to the appropriate
  scale prescribed by the Secretary of State.

Audit Commission Act 1998 - 1998 CHAPTER 18 -
http//www.opsi.gov.uk/acts/acts1998/ukpga_1998001
8_en_2
10
Seen in context..

• membrane of vacuole
• with cell cycle-independent morphology
• def The lipid bilayer surrounding a vacuole that
  retains the same shape regardless of cell cycle
  phase. The membrane separates its contents from
  the cytoplasm of the cell.

membrane of vacuole with cell cycle-correlated
morphology def The lipid bilayer surrounding a
vacuole, the shape of which correlates with cell
cycle phase. The membrane separates its contents
from the cytoplasm of the cell.
11
With some inside information
Plant vacuole membrane

• membrane of vacuole
• with cell cycle-independent morphology
• def The lipid bilayer surrounding a vacuole that
  retains the same shape regardless of cell cycle
  phase. The membrane separates its contents from
  the cytoplasm of the cell.
• membrane of vacuole
• with cell cycle-correlated morphology
• def The lipid bilayer surrounding a vacuole, the
  shape of which correlates with cell cycle phase.
  The membrane separates its contents from the
  cytoplasm of the cell.

Fungal vacuole membrane
12
Error checking
!?
13
The Proposed Solution

• We can apply taxon information in a rigorous and
  non-misleading way using relationships
• is_relevant_for
• is_only_in
• applies_to_all

14

• Relevance
• is_relevant_to

15
Relevance is_relevant_to

• This GO term can be used for annotation of this
  taxon.
• The GO term may not be appropriate for all
  subgroups of this taxon.
• The GO term may not be appropriate for all
  supergroups of this taxon.

16

• membrane of
• vacuole
• with cell cycle
• -independent
• morphology

is_relevant_for
Arabidopsis
membrane of vacuole with cell cycle- correlated
morphology
Saccharomyces
is_relevant_for
17
Relevance relationships

• Added by curators.
• Inferred from annotations.

18
Relevance - Propagation

• Propagates up the GO.
• Propagates up the taxon hierarchy.

19
Relevance
20
Relevance
21
Relevance
cellular organisms Eukaryota Fungi/Metazoa
group Metazoa Eumetazoa Bilateria
Coelomata Deuterostomia Chordata Craniata
Vertebrata Gnathostomata Teleostomi
Euteleostomi Actinopterygii Actinopteri
Neopterygii Teleostei Elopocephala
Clupeocephala Otocephala Ostariophysi
Otophysi Cypriniphysi Cypriniformes
Cyprinoidea Cyprinidae Rasborinae Danio
22
Relevance
cellular organisms Eukaryota Fungi/Metazoa
group Metazoa Eumetazoa Bilateria
Coelomata Deuterostomia Chordata Craniata
Vertebrata Gnathostomata Teleostomi
Euteleostomi Actinopterygii Actinopteri
Neopterygii Teleostei Elopocephala
Clupeocephala Otocephala Ostariophysi
Otophysi Cypriniphysi Cypriniformes
Cyprinoidea Cyprinidae Rasborinae Danio
23

• Specificity
• is_only_in

24
Specificity is_only_in

• The GO term never applies to species outside of
  this taxon.
• (The GO terms may apply to only some subtaxa of
  the taxon, and not all.)

25
Viridiplantae
http//www.seaweed.ie/algae/images/hydrodictyon2.j
pg
http//www.blackburn.gov.uk/server.php?showconMed
iaFile.7355
green plants and algae
26
http//www.blackburn.gov.uk/server.php?showconMed
iaFile.7355
Leaf development never occurs outside of
Viridiplantae.
Not all subtaxa of Viridiplantae make leaves.
http//www.seaweed.ie/algae/images/hydrodictyon2.j
pg
27
formally

• GO term leaf development
• relationship is_only_in
• taxon Viridiplantae
• Leaf development never occurs outside of
  Viridiplantae.
• (Not all subtaxa of Viridiplantae make leaves.)

28
Specificity - is_only_in

• Added manually by curators as they develop the
  graph.

29
Specificity - is_only_in
GO Taxon

• Propagates down the GO.
• Propagates up the taxon hierarchy.

30
Specificity - is_only_in
31
Specificity - is_only_in
32
Specificity - is_only_in
Cellular organisms Eukaryota
Viridiplantae

33
Specificity - is_only_in
Cellular organisms Eukaryota
Viridiplantae

34
What about convergent evolution?
http//www.dcn.davis.ca.us/go/platypus/images/plat
ygeo.jpg
http//homepage.mac.com/wildlifeweb/bird/mallard/m
allard_male_image_11tfk.jpg
35
Assumptions- beak development

• These beaks were produced by convergent evolution
• The two beak development processes are the same.

36
one beak development term

• is_relevant_for Anas platyrhynchos
• (mallard)
• is_relevant_for Ornithorhynchus anatinus
• (Platypus)
• is_only_in Amniota
• (or vertebrata if that is easier)

37

• Validity
• applies_to_all

38
Validity applies_to_all

• The GO term applies to all subtaxa of a given
  taxon.
• (The GO term may also apply to species outside of
  the taxon.)

39
Validity applies_to_all
Homoiothermy is exhibited by all subtaxa of
the taxon Mammalia.
http//www.thermalconsultancy.com/
http//coolcosmos.ipac.caltech.edu/image_galleries
/ir_zoo/coldwarm.html
40
Validity applies_to_all
Homoiothermy is absent in many of the taxonomic
groups outside of Mammalia.
http//coolcosmos.ipac.caltech.edu/image_galleries
/ir_zoo/coldwarm.html
41
Validity applies_to_all
with some exceptions
Homoiothermy is also present in birds, tuna fish,
and honey bees.
Bee Hive
http//www.beebehavior.com/
42
formally

• GO term homoiothermy
• relationship applies_to_all
• taxon Mammalia
• Homoiothermy is found in all subtaxa of the taxon
  Mammalia.
• (Homoiothermy is also found outside of the taxon
  Mammalia.)

43
Validity

• Propagates up the GO.
• Propagates down the taxon hierarchy.

44
Validity
45
Validity
46
Validity
47
Validity applies_to_all

• Added manually by curators

48
Cave fish

• Not all fish carry out eye development processes.

http//www.petpig.com/fish/fish/blind_cave_fish1.j
pg
49

• fish

50

• fish

eye development applies_to_all
?
51
fish
eye development applies_to_all

• other
• fish
• cave other
• fish fish

52
fish

• other
• fish
• cave other
• fish fish

53
What does this buy us?
54
Users no longer hunting for terms in the GO
haystack.
http//bestblog.mlblogs.com/inside_pitch/images/ne
edle_in_the_haystack_4.jpg
55
All species.
56
Mammal-specific terms greyed out.
57
All species.
58
Insect-specific terms greyed out.
59
Clues to language

• membrane of vacuole
• with cell cycle-independent morphology
• def The lipid bilayer surrounding a vacuole that
  retains the same shape regardless of cell cycle
  phase. The membrane separates its contents from
  the cytoplasm of the cell.

60
Clues to language

• membrane of vacuole
• with cell cycle-independent morphology
• def The lipid bilayer surrounding a vacuole that
  retains the same shape regardless of cell cycle
  phase. The membrane separates its contents from
  the cytoplasm of the cell.
• is_relevant_for Viridiplantae

61
Automatic Annotation and Ontology Error Checking
Brown Arabidopsis annotations. Blue
Drosophila annotations.
62
Brown Arabidopsis annotations. Blue
Drosophila annotations.
63
Brown Arabidopsis annotations found. Blue
Drosophila annotations found.
64
Error checking
65
GO-Taxon relationships

• To be put in an external OBO format file.
• GO file link
  file slimmed
• 
  taxonomy
  file

66
Acknowledgements

• Waclaw Kusnierczyk.
• Taxonomy-based partitioning of the Gene Ontology
• In press Journal of Biomedical Informatics.
• Proposal by
• Jennifer Deegan
• Chris Mungall
• Jane Lomax

With input from Midori Harris Melissa
Haendel Judy Blake John Day-Richter Waclaw
Kusnierczyk David Hill