GO Centric View

[karenross]

The GO-centric view does not show GO Molecular Function terms (Biological Process and Cellular Component terms are OK). For example, see https://proconsortium.org/app/entry/PR:000025934/

[Julie-Cowart]

Turns out it is literally hard coded

        if r['relation'] == 'has_function':
            data['go']['Molecular Function'][(r['ontology_id'], modifier)].append(r)
        elif r['relation'] == 'participates_in':
            data['go']['Biological Process'][(r['ontology_id'], modifier)].append(r)
        elif r['relation'] == 'located_in':
            data['go']['Cellular Location'][(r['ontology_id'], modifier)].append(r)

Problem is that we seem to be using 'capable_of' now instead of has_function. I can fix by changing the code here.
Are there any other relations we are missing right now that should be included under one of these headings? Here are the relations found in the paf as of the last release.

Relation	count
participates_in	1687
capable_of	1554
located_in	1100
has_part	616
associated_with_disease_progression	606
part_of	97
has_agent	27
has_role	10
associated_with_disease_suppression	4
expressed_in	4

[Julie-Cowart]

Here is how they seem to be used

has_part -> Pfam
associated_with_disease_progression -> Disease Ontology
part_of -> GO complex
has_agent -> Sequence ontology
has_role -> CHEBI:59163 biomarker
associated_with_disease_suppression -> Disease Ontology
expressed_in -> Cell ontology

So the only one that is using GO terms is part_of

[nataled]

In addition to the relations used in PAF, we also have some used in the OBO. I'll list them here for completeness, then organize them based on function, process, or component considerations in a later comment. Before I do so, however, one question I have is whether or not the annotations shown in the table must come exclusively from the PAF, or does the code allow for these to come from the OBO? If the former, then I might have to move some to the PAF.

Relation	Count
only_in_taxon	74826
has_canonical_sequence	38052
has_gene_template	37959
has_part	23479
lacks_part	6494
has_component	1639
output_of	809
derives_from	57
non-covalently_bound_to	41
confers_resistance_to	25
capable_of	25
part_of	19
involved_in	3
located_in	2
has_quality	1

[Julie-Cowart]

It currently only shows the PAF

[nataled]

Okay, we can keep it like that and move any relevant annotations that are currently only in OBO to also appear in PAF. I guess the question then becomes: What do we do with annotations that don't use GO (such as the disease-related annotations)? Right now those don't appear at all. Let's discuss that on the PRO call this Friday.

[nataled]

My first pass categorization. Note:

1. Some relations are used for definitions in OBO but can be also used for annotation in PAF.
2. Some complications noted.
3. Possibility, should we decide to do more with the GO-centric view (unlikely), is to rename the following in the GO-centric view:

• 'Molecular Function GO Term' to 'Capability Annotations'
• 'Biological Process GO Term' to 'Process Annotations'
• 'Cellular Location GO Term' to 'Location Annotations'

4. Following from 3 above, rename the whole table from 'Functional Annotation' to 'Annotations', rename 'PRO Centric View' to 'Entity-centric View', and rename 'GO Centric view' to 'Annotation-centric View'

Again, to be discussed.

Relation	Def	F	P	C	Other	Note
participates_in (PAF)			x			GO: Participation that is not necessarily enabling
involved_in (OBO)	x		x			GO: Participation that is enabling
capable_of (PAF)		x				GO:
capable_of (OBO)	x	x				GO:
located_in (PAF)				x		GO:
located_in (OBO)	x			x		GO:
has_part (PAF)					x	Pfam:
has_part (OBO)	x					various
associated_with_disease_progression			x			DOID:
associated_with_disease_suppression			x			DOID:
part_of (PAF)		x				GO: (complexes) Technically component, but in PAF it's more a functional annotation
part_of (OBO)	x					various; In OBO these are used for definitions
has_agent					x	SO: Used to indicate the type of mutation
has_role		x				CHEBI:
expressed_in				x		CL:
only_in_taxon	x					NCBITaxon:
has_canonical_sequence					x	UniProtKB:
has_gene_template	x					various
lacks_part	x					various
has_component	x					various
output_of	x		x			GO:
derives_from	x					PR:
non-covalently_bound_to	x					CHEBI:
confers_resistance_to		x				CHEBI:
has_quality	x					SO: