The SimpleKnowledge Editor is a tool to describe straightforward ontologies that do not involve reasoning, based solely on assertion triplets. The SimpleKnowledge Editor is thus an alternative to OWL editors such as Stanford's Protege.
The SimpleKnowledge Editor is a tool to build the source for a RDF knowledge graph. A knowledge graph can be considered to be a set of assertions. An assertion defines a relationship between two entities.
In a graph database, an assertion is represented by two nodes (representing the originating and terminating entities) linked by an edge (representing the relationship). The combination of two nodes and a relationship is known as a triplet.
In graph databases such as neo4j, both nodes and edges can have attributes, or properties. A graph that contains assertions with nodes, edges, and properties is known as a property graph:
A graph that represents properties as nodes is known as a RDF graph. For example, the properties property 1 = A and property 2 = B can be represented with nodes that are themselves in a hierarchy:
The SimpleKnowledge Editor is a Google Sheets solution to generate a RDF knowledge graph. The primary application of the SimpleKnowledge Editor is the creation of source that can be imported and translated into an instance of a Unified Biological Knowledge Graph.
The UBKG graph architecture is based on the UMLS model, in which a concept in a vocabulary (e.g., a SimpleKnowledge ontology) is represented with
- a Concept node for the unique concept in the knowledge graph
- a Code node that the vocabulary uses to represent the concept
- a Preferred Term--a Term node of type PT
- one or more Synonyms-- Term nodes of type SY
- one or more Definitions-- Definition nodes
The UBKG also allows concept cross-referencing (also known as equivalence classes), in which Code nodes share links to Concept nodes. When the code in a vocabulary is cross-referenced to a code in another vocabulary, the Code node is linked to the Concept node associated with the other Code. This allows for code synonymy by traversal across vocabularies.
The following illustrates how "right kidney" is encoded in the UBKG in multiple vocabularies. The codes in the vocabularies share concept C0227613:
| Code | Term |
|---|---|
| SENNET:C030030 | Kidney (Right) |
| HUBMAP:C030059 | Kidney (Right) |
| NCI:C34005 | Right Kidney |
| UWDA:7204 | Right kidney |
| SNOMEDCT_US:9846003 | Right kidney structure |
| UBERON:0004539 | right kidney |
| CHV:0000022146 | right kidney |
To build a SimpleKnowledge Editor for a knowledge graph, make a copy of the Example SimpleKnowledge Editor document. Work in the sheet named SimpleKnowledge Editor.
The cells in the SimpleKnowledge Editor sheet represent the nodes and edges of the knowledge graph. The SimpleKnowledge Editor uses data validation rules to enforce the required structure of the nodes and edges.
To define a node, complete columns A-F.
This is the required preferred term for the node. The term must be unique.
This is the code for the node in the ontology represented by the SimpleKnowledge Editor. The format of the code should be SAB:CODE, in which SAB is an acronym for ontology. Codes must be unique.
This is an optional definition for the code.
This is an optional list of synonyms for the code. The list should be pipe-delimited.
Identifier for the optional equivalence class or cross-reference for the code. Identifiers can be in format:
- Concept Unique Identifier
- Code in another vocabulary
| term | code | definition | synonym | dbxref |
|---|---|---|---|---|
| Color | ABC:C001 | visual perception based on the electromagnetic spectrum | hue|cast|tint | UMLS:C0009393 |
Column F defines isa, or hierchical, edges between code nodes. A code can be in multiple isa hierarchies. To define multiple hierarchies, use comma delimiting in the isa column. Identify a node for an isa relationship using the term in column A, not the code in colum B.
The ABC:C003 (blue) code is a member of two hierarchies--one with parent ABC:C001 (Color) and one with parent ABC:C002 (Emotion).
| term | code | definition | synonym | dbxref | isa |
|---|---|---|---|---|---|
| Color | ABC:C001 | visual perception based on the electromagnetic spectrum | hue|cast|tint | UMLS:C0009393 | |
| Emotion | ABC:C002 | physical and mental states brought on by neurophysiological changess | feeling | ||
| blue | ABC:C003 | Color,Emotion |
- There must be a root node with no value in the isa column.
- All other nodes must have at least one reference in the isa column.
- Values in isa must be terms in Column A.
The columns after F can be used to define custom edges. The cell that intersects a node row and an edge column defines the edge between the node's row and other nodes.
Note:
- The thing root node.
- The Color, Emotion, and sky nodes are all members of the thing hierarchy.
- The blue node is a member of both the Color and Emotion hierarchies.
- The cyan and azure nodes are children of blue, and thus are in the Color hierarchy.
- The sky node is used in the used_to_describe custom edge column. This represents the assertion azure used_to_describe sky.
| term | code | definition | synonym | dbxref | isa | used_to_describe |
|---|---|---|---|---|---|---|
| thing | ABC:C000 | root | ||||
| Color | ABC:C001 | visual perception based on the electromagnetic spectrum | hue|cast|tint | UMLS:C0009393 | thing | |
| Emotion | ABC:C002 | physical and mental states brought on by neurophysiological changess | feeling | thing | ||
| blue | ABC:C003 | Color,Emotion | ||||
| cyan | ABC:C004 | blue | ||||
| azure | ABC:C005 | blue | sky | |||
| sky | ABC:C006 | thing |
This will result in a graph similar to
It is possible to visualize parts of a SimpleKnowledge knowledge graph using the custom guesdt application. guesdt is a Javascript app that works with the SimpleKnowledge source in Google Sheet.
To visualize a SimpleKnowledge graph, enter into a browser a URL in format https://guesdt.com/SimpleKnowledgeVisualization.html?sheet=sheetid.
Obtain the id for the Google Sheet used by the guesdt from the URL that points to the Google Sheet--e.g.,
The following image is from the visualization of the example SimpleKnowledge graph. Note that the visualization automatically creates inverse relationships (e.g., inverse_isa).
