+ Multi-organism GO annotation David Osumi-Sutherland Gene Ontology
Jan 02, 2016
+
“rif... [is] expressed on the infected red [blood] cell surface.”
Plasmodium falciparum
GP of one organism in CC of another
+
Plasmodium falciparum| Homo sapiens
• The species in the interaction are recorded... by entering two taxon IDs in the Taxon column:• First taxon ID: species encoding the gene product• First taxon ID: other species in the interaction.
GP of one organism in CC of another
+
Plasmodium falciparum| Homo sapiens
• The species in the interaction are recorded... by entering two taxon IDs in the Taxon column:• First taxon ID: species encoding the gene product• First taxon ID: other species in the interaction.
Why not just plasma membrane?
GP of one organism in CC of another
+A duplicate cell component branch is growing under host-cell
Why? Keeps annotations of gene
products located in CCs of host organism separate from regular GO CC annotations.
+A duplicate cell component branch is growing under host-cell
Many terms duplicate the regular cell component hierarchy. THIS IS NOT SUSTAINABLE
Some terms refer to components that are formed from the interaction of 2 organisms. Add new parent class for
these: multi-organism cell component
+Proposed solution
Plasmodium falciparumparasite_of Homo sapiens
located_in
GO:0005886 plasma membrane
+ annotation extension: part_of Homo sapiens
+Proposed solution
Plasmodium falciparumparasite_of Homo sapiens
located_in
GO:0005886 plasma membrane
• When used, replaces part_of as default GP to CC relation in GPAD• Tools used for grouping annotations need to group these annotations
separately. Precedent: NOT
New qualifier / GP to CC relation: located_in
+ annotation extension: part_of Homo sapiens
+Proposed solution
Plasmodium falciparumparasite_of Homo sapiens
located_in
GO:0005886 plasma membrane
New qualifier / GP to CC relation: located_in
+ annotation extension: part_of Homo sapiens
Replace host CC term with regular one, recording the taxon of the organism it is part of in the annotation extension.
+Proposed solution
Plasmodium falciparumparasite_of Homo sapiens
located_in
GO:0005886 plasma membrane
Taxon extension – using standard relations between organisms :
host_of, symbiont_of, parasite_of vector_for…
Relations defined with population and community ontology (PCO)
+ annotation extension: part_of Homo sapiens
+Proposed solution – retrofit when host not recorded
Plasmodium falciparumsymbiont_of host organism *
located_in
GO:0005886 plasma membrane
+ annotation extension: part_of host organism *
* Generic host organism class from PCO
+
Homo sapiens host_ofHuman rhinovirus 3
Proposed solution
+ annotation extension: occurs_in Homo sapiens
Optionally record where the process takes place
+
mate_of
host_of, parasite_of, vector_for
host_of/parasite_of
bitten_by ?
pregnant_mother_of ?/ fetus_of
?
+What does annotation mean?
Too strong: All instances of the RIF gene product are located in some
plasma membrane that is part of some Homo sapiens. All instances of the ICAM-1 gene product are involved in
some HRV3 virion attachment to (human) host cell.
+Capabilities are not necessarily realised
All sperm capable_of (some) fertilization But not all sperm do. In fact most do not.
All ICAM-1 gene product capable_of_part_of some ‘virion attachment to host cell’ But most don’t.
+LEGO-ish class level formalisation of ‘lego’ brick
Gask3 SubClassOf: capable_of some ('protein kinase activity'
that (part_of some 'canonical wnt signalling’)and (occurs_in some cytosol))
protein kinase activity
canonical Wnt signaling
P
Gask3b
cytosol
+The same pattern works for three separate annotations
Gask3 SubClassof:capable_of some 'protein kinase activity’
capable_of some('molecular_function' that (part_of some 'canonical wnt signalling'))
capable_of some ('molecular_function' that (occurs_in some cytosol))
protein kinase activity
canonical Wnt signaling
P
Gask3b
cytosol
+Using this pattern for localisation to CC of an interacting organismRIF subClassOf: capable_of some ( 'molecular function' that occurs_in_location some ( 'plasma membrane' that part_of some ( 'Homo sapiens' that host_of some 'Plasmodium falciparum 3D7’ ) ))
occurs_in_location = more general version of occurs_in that entails location rather than parthod
+Using this pattern for multi-organism process
ICAM1 subClassOf: capable_of some ( 'molecular function' that part_of some ( 'receptor-mediated virion attachment to host cell' that occurs_in some ( ‘Homo sapiens' that host_of some 'Human rhinovirus 3' ) ))
ICAM1 encoded_bysome Homo sapiens
+
Ba71V-98 subClassOf: capable_of some ( 'molecular function' that part_of some ( 'virion attachment to host cell' that occurs_in some ( Cercopithecus aethiops that host_of some 'African swine fever virus’ ) ))
Ba71V-98 encoded_by some‘African swine fever virus’
ASFV virion
Using this pattern for multi-organism process
+Multi-organism lego brick
ICAM1 subClassOf: capable_of some ( ’protein binding' that part_of some ( 'receptor-mediated virion attachment to host cell' that occurs_in some ( ‘plasma membrane’ that part_of some ( ‘Homo sapiens' that host_of some 'Human rhinovirus 3’ ) ) ))
ICAM1 encoded_bysome Homo sapiens
+But this pattern can’t work directly for Lego construction
Visualization
protein kinase activity
canonical Wnt signaling
P
Gask3b
cytosol
protein binding
Btrc
cytosol
required_for
Individual: gobp-0000000001Types: ‘cannonical wnt signaling’#Individual: gomf-0000012345Types: ‘protein kinase activity occurs_in some cytosol, enabled_by some PR:P87654 Facts: required_for gomf-0000012346Individual: gomf-0000012346Types: ‘protein binding’, occurs_in some cytosol, enabled_by some Btrc
+Pattern is incompatible with current Lego pattern
Visualization
protein kinase activity
canonical Wnt signaling
P
Gask3b
cytosol
protein binding
Btrc
cytosol
required_for
Individual: gobp-0000000001Types: ‘cannonical wnt signaling’#Individual: gomf-0000012345Types: ‘protein kinase activity occurs_in some cytosol, enabled_by some PR:P87654 Facts: required_for gomf-0000012346Individual: gomf-0000012346Types: protein binding, occurs_in some cytosol, enabled_by some Btrc
Note – CC & GP are classes
+Possible Solutions:
Parallel translations? Same basic pattern used in each case Lego bricks capture a set of related findings in a single
paper Class level assertion makes sense for this Lego construction captures a model of how these fit
together under some circumstances…
Change LEGO pattern CC & GP both become individuals