Skip to content

Latest commit

 

History

History
125 lines (88 loc) · 14.1 KB

README.md

File metadata and controls

125 lines (88 loc) · 14.1 KB

Knowledge Graph Construction Approaches

Wiki Page: KG-Construction
Jupyter Notebook: Data_Preparation.ipynb
Required Input: resources/construction_approach/subclass_construction_map.pkl

Purpose

New data can be added to the knowledge graph using 2 different construction approaches: (1) instance-based or (2) subclass-based. Each of these approaches is described further below.


Instance-Based Construction

In this approach, each new edge is added as an instance of an existing class (via rdf:Type) in the knowledge graph.

EXAMPLE: Adding the edge: Morphine ➞ isSubstanceThatTreats ➞ Migraine

  • isSubstanceThatTreats(Morphine, x1)
  • Type(x1, Migraine)

In this example, Morphine is an ontology data node from ChEBI and Migraine is a Human Phenotype Ontology term. This would result in the following triples:

UUID1 = MD5(Morphine + isSubstanceThatTreats + Migraine + "subject")
UUID2 = MD5(Morphine + isSubstanceThatTreats + Migraine + "object")

UUID1, rdf:type, Morphine
UUID1, rdf:type, owl:NamedIndividual

UUID2, rdf:type, Migraine
UUID2, rdf:type, owl:NamedIndividual

UUID1, isSubstanceThatTreats, UUID2

A table is provided below showing the different triples that are added as function of edge type (i.e. class-class vs. class-instance vs. instance-instance) and relation strategy (i.e. relations only or relations + inverse relations).

Edge Type Relations Needed Triples
Class-Class Relations Only GO_1234567, REL, DOID_1234567

UUID1 = pkt + MD5(GO_1234567 + <<REL>> + DOID_1234567 + "subject")
UUID2 = pkt + MD5(GO_1234567 + <<REL>> + DOID_1234567 + "object")
UUID1, rdf:type, GO_1234567
UUID1, rdf:type, owl:NamedIndividual

UUID2, rdf:type, DOID_1234567
UUID2, rdf:type, owl:NamedIndividual

UUID1, REL, UUID2
Class-Class Relations + Inverse Relations GO_1234567, REL, DOID_1234567
DOID_1234567, INV_REL, GO_1234567

UUID1 = pkt + MD5(GO_1234567 + <<REL>> + DOID_1234567 + "subject")
UUID2 = pkt + MD5(GO_1234567 + <<REL>> + DOID_1234567 + "object")
UUID1, rdf:type, GO_1234567
UUID1, rdf:type, owl:NamedIndividual

UUID2, rdf:type, DOID_1234567
UUID2, rdf:type, owl:NamedIndividual

UUID1, REL, UUID2
UUID2, INV_REL, UUID1
Class-Instance Relations Only GO_1234567, REL, HGNC_1234567

UUID1 = pkt + MD5(GO_1234567 + <<REL>> + HGNC_1234567 + "subject")
UUID2 = pkt + MD5(GO_1234567 + <<REL>> + HGNC_1234567 + "object")
UUID1, rdf:type, GO_1234567
UUID1, rdf:type, owl:NamedIndividual

HGNC_1234567, rdfs:subClassOf, subclass_dict[HGNC_1234567]
HGNC_1234567, rdf:type, owl:Class
UUID2, rdf:type, HGNC_1234567
UUID2, rdf:type, owl:NamedIndividual

UUID1, REL, UUID2
Class-Instance Relations + Inverse Relations GO_1234567, REL, HGNC_1234567
HGNC_1234567, INV_REL, GO_1234567

UUID1 = pkt + MD5(GO_1234567 + <<REL>> + HGNC_1234567 + "subject")
UUID2 = pkt + MD5(GO_1234567 + <<REL>> + HGNC_1234567 + "object")
UUID1, rdf:type, GO_1234567
UUID1, rdf:type, owl:NamedIndividual

HGNC_1234567, rdfs:subClassOf, subclass_dict[HGNC_1234567]
HGNC_1234567, rdf:type, owl:Class
UUID2, rdf:type, HGNC_1234567
UUID2, rdf:type, owl:NamedIndividual

UUID1, REL, UUID2
UUID2, INV_REL, UUID1
Instance-Instance Relations Only HGNC_1234567, REL, HGNC_7654321

UUID1 = pkt + MD5(HGNC_1234567 + <<REL>> + HGNC_7654321 + "subject")
UUID2 = pkt + MD5(HGNC_1234567 + <<REL>> + HGNC_7654321 + "object")
HGNC_1234567, rdfs:subClassOf,subclass_dict[HGNC_1234567]
HGNC_1234567, rdf:type, owl:Class
UUID1, rdf:type, HGNC_1234567
UUID1, rdf:type, owl:NamedIndividual

HGNC_7654321, rdfs:subClassOf,subclass_dict[HGNC_7654321]
HGNC_7654321, rdf:type, owl:Class
UUID2, rdf:type, HGNC_7654321
UUID2, rdf:type, owl:NamedIndividual

UUID1, REL, UUID2
Instance-Instance Relations + Inverse Relations HGNC_1234567, REL, HGNC_7654321
HGNC_7654321, INV_REL, HGNC_1234567

UUID1 = pkt + MD5(HGNC_1234567 + <<REL>> + HGNC_7654321 + "subject")
UUID2 = pkt + MD5(HGNC_1234567 + <<REL>> + HGNC_7654321 + "object")
HGNC_1234567, rdfs:subClassOf,subclass_dict[HGNC_1234567]
HGNC_1234567, rdf:type, owl:Class
UUID1, rdf:type, HGNC_1234567
UUID1, rdf:type, owl:NamedIndividual

HGNC_7654321, rdfs:subClassOf,subclass_dict[HGNC_7654321]
HGNC_7654321, rdf:type, owl:Class
UUID2, rdf:type,HGNC_7654321
UUID2, rdf:type,owl:NamedIndividual

UUID1, REL, UUID2
UUID2, INV_REL, UUID1

Note. UUID is a BNode that is created from an md5 hash of concatenated URIs. The URIs used in the hash string includes the subject and object URIs (each appended with "subject" and "object", respectively) in addition to a relation (<<REL>>). To account for future use cases, we have devised a heuristic to determine what is used for <<REL>>: (1) for a given relation, determine if it has an inverse (via owl:InverseOf); (2) sort the relations alphabetically; and (3) select the first relation. The selected relation is then used for creating both UUID BNodes (i.e. UUID1 and UUID2 in all examples in the table above). For example, if the relations were causes and caused_by both UUID BNodes would be created using caused_by. Please note that all UUID BNodes created during the construction process are explicitly defined within the pkt namespace (https://github.com/callahantiff/PheKnowLator/pkt/).



Subclass-Based Construction

In this approach, each new edge is added as a subclass of an existing ontology class (via rdfs:subClassOf) in the knowledge graph.

EXAMPLE: Adding the edge: TGFB1 ➞ participatesIn ➞ Influenza Virus Induced Apoptosis

  • participatesIn(TGFB1, Influenza Virus Induced Apoptosis)
  • subClassOf(Influenza Virus Induced Apoptosis, Influenza A Pathway)
  • Type(Influenza Virus Induced Apoptosis, owl:Class)

Where TGFB1 is a Protein Ontology term and Influenza Virus Induced Apoptosis is a non-ontology data node from Reactome. In this example, Influenza A Pathway is an existing Pathway Ontology class. This would result in the following triples:

UUID1 = MD5(TGFB1 + participatesIn + Influenza Virus Induced Apoptosis)
UUID2 = MD5(TGFB1 + participatesIn + Influenza Virus Induced Apoptosis + owl:Restriction)

Influenza Virus Induced Apoptosis, rdfs:subClassOf, Influenza A Pathway
Influenza Virus Induced Apoptosis, rdf:type, owl:Class

UUID1, rdfs:subClassOf, TGFB1
UUID1, rdfs:subClassOf, UUID2
UUID2, rdf:type, owl:Restriction
UUID2, owl:someValuesFrom, Influenza Virus Induced Apoptosis
UUID2, owl:onProperty, participatesIn

A table is provided below showing the different triples that are added as function of edge type (i.e. class-class vs. class-instance vs. instance-instance) and relation strategy (i.e. relations only or relations + inverse relations).

Edge Type Relations Needed Triples
Class-Class Relations Only GO_1234567, REL, DOID_1234567

UUID1 = pkt + MD5(DOID_1234567 + REL + GO_1234567)
UUID2 = MD5(DOID_1234567 + REL + GO_1234567 + owl:Restriction)
UUID1, rdfs:subClassOf, GO_1234567
UUID1, rdfs:subClassOf, UUID2
UUID2, rdf:type, owl:Restriction
UUID2, owl:someValuesFrom, DOID_1234567
UUID2, owl:onProperty, REL
Class-Class Relations + Inverse Relations GO_1234567, REL, DOID_1234567
DOID_1234567, INV_REL, GO_1234567

UUID1 = pkt + MD5(GO_1234567 + REL + DOID_1234567)
UUID2 = MD5(GO_1234567 + REL + DOID_1234567 + owl:Restriction)
UUID3 = pkt + MD5(DOID_1234567 + INV_REL + GO_1234567)
UUID4 = MD5(DOID_1234567 + INV_REL + GO_1234567 + owl:Restriction)
UUID1,rdfs:subClassOf,GO_1234567
UUID1,rdfs:subClassOf,UUID2
UUID2, rdf:type, owl:Restriction
UUID2, owl:someValuesFrom,DOID_1234567
UUID2, owl:onProperty, REL

UUID3, rdfs:subClassOf,DOID_1234567
UUID3,rdfs:subClassOf,UUID4
UUID4, rdf:type, owl:Restriction
UUID4, owl:someValuesFrom, GO_1234567
UUID4, owl:onProperty, INV_REL
Class-Instance Relations Only GO_1234567, REL, HGNC_1234567

UUID1 = pkt + MD5(GO_1234567 + REL + HGNC_1234567)
UUID2 = MD5(GO_1234567 + REL + HGNC_1234567 + owl:Restriction)
HGNC_1234567, rdfs:subClassOf, subclass_dict[HGNC_1234567]
HGNC_1234567, rdf:type, owl:Class

UUID1, rdfs:subClassOf, GO_12334567
UUID1, rdfs:subClassOf, UUID2
UUID2, rdf:type, owl:Restriction
UUID2, owl:someValuesFrom, HGNC_1234567
UUID2, owl:onProperty, REL
Class-Instance Relations + Inverse Relations GO_1234567, REL, HGNC_1234567
HGNC_1234567, INV_REL, GO_1234567

UUID1 = pkt + MD5(GO_1234567 + REL + HGNC_1234567)
UUID2 = MD5(GO_1234567 + REL + HGNC_1234567 + owl:Restriction)
UUID3 = pkt + MD5(HGNC_1234567 + INV_REL + GO_1234567)
UUID4 = MD5(HGNC_1234567 + INV_REL + GO_1234567 + owl:Restriction)
HGNC_1234567, rdfs:subClassOf, subclass_dict[HGNC_1234567]
HGNC_1234567, rdf:type, owl:Class

UUID1, rdfs:subClassOf, GO_12334567
UUID1, rdfs:subClassOf, UUID2
UUID2, rdf:type, owl:Restriction
UUID2, owl:someValuesFrom, HGNC_1234567
UUID2, owl:onProperty, REL

UUID3, rdfs:subClassOf, HGNC_1234567
UUID3, rdfs:subClassOf, UUID4
UUID4, rdf:type, owl:Restriction
UUID4, owl:someValuesFrom, GO_12334567
UUID4, owl:onProperty, INV_REL
Instance-Instance Relations Only HGNC_1234567, REL, HGNC_7654321

UUID1 = pkt + MD5(HGNC_1234567 + REL + HGNC_7654321)
UUID2 = MD5(HGNC_1234567 + REL + HGNC_7654321 + owl:Restriction)
HGNC_1234567, rdfs:subClassOf, subclass_dict[HGNC_1234567]
HGNC_1234567, rdf:type, owl:Class

HGNC_7654321, rdfs:subClassOf, subclass_dict[HGNC_7654321]
HGNC_7654321, rdf:type, owl:Class

UUID1, rdfs:subClassOf, HGNC_1234567
UUID1, rdfs:subClassOf, UUID2
UUID2, rdf:type, owl:Restriction
UUID2, owl:someValuesFrom, HGNC_7654321
UUID2, owl:onProperty, REL
Instance-Instance Relations + Inverse Relations HGNC_1234567, REL, HGNC_7654321
HGNC_7654321, INV_REL, HGNC_1234567

UUID1 = pkt + MD5(HGNC_1234567 + REL + HGNC_7654321)
UUID2 = MD5(HGNC_1234567 + REL + HGNC_7654321 + owl:Restriction)
UUID3 = pkt + MD5(HGNC_7654321 + INV_REL + HGNC_1234567)
UUID4 = MD5(HGNC_7654321 + INV_REL + HGNC_1234567 + owl:Restriction)
HGNC_1234567, rdfs:subClassOf, subclass_dict[HGNC_1234567]
HGNC_1234567, rdf:type, owl:Class

HGNC_7654321, rdfs:subClassOf, subclass_dict[HGNC_7654321]
HGNC_7654321, rdf:type, owl:Class

UUID1, rdfs:subClassOf, HGNC_1234567
UUID1, rdfs:subClassOf, UUID2
UUID2, rdf:type, owl:Restriction
UUID2, owl:someValuesFrom, HGNC_7654321
UUID2, owl:onProperty, REL

UUID3, rdfs:subClassOf, HGNC_7654321
UUID3, rdfs:subClassOf, UUID4
UUID4, rdf:type, owl:Restriction
UUID4, owl:someValuesFrom, HGNC_1234567
UUID4, owl:onProperty, INV_REL

Note. When UUID is used within the subclass-based construction approach it is a BNode that is created from an md5 hash of concatenated URIs. See each table row for a sample of what this looks like. Please note that the primary UUID BNodes created during the construction process are explicitly defined within the pkt namespace (https://github.com/callahantiff/PheKnowLator/pkt/).


🛑 ASSUMPTIONS 🛑
Data: subclass_construction_map.pkl

The algorithm makes the following assumptions:

  • Make sure that you have created the non-ontology node data to ontology class mapping dictionary (described below) to the ./resources/construction_approach/*.pkl directory.

Input requirements for both approaches: A pickled dictionary (keys contain node identifiers (non-ontology node data) and the values are lists of ontology class identifiers) added to the ./resources/construction_approach/ directory. An example of this dictionary is shown below:

{
  'R-HSA-168277'  : ['http://purl.obolibrary.org/obo/PW_0001054',
                     'http://purl.obolibrary.org/obo/GO_0046730'],
  'R-HSA-9026286' : ['http://purl.obolibrary.org/obo/PW_000000001',
                     'http://purl.obolibrary.org/obo/GO_0019372'],
  '100129357'     : ['SO_0000043'],
  '100129358'     : ['SO_0000336'],
}

Please see the Reactome Pathways - Pathway Ontology and Genomic Identifiers - Sequence Ontology sections of the Data_Preparation.ipynb Jupyter Notebook for examples of how to construct this document.