Proposition

Trial Use

May change in future releases. See Maturity Model.

Abstract Class — not instantiated directly; concrete subclasses inherit its attributes.

Computational Definition

An abstract entity representing a possible fact that may be true or false. As abstract entities, Propositions capture a ‘sharable’ piece of meaning whose identify and existence is independent of space and time, or whether it is ever asserted to be true by some agent.

Information Model

Some Proposition attributes are inherited from Entity.

Field	Flags	Type	Limits	Description
id		string	0..1	The ‘logical’ identifier of the Entity in the system of record, e.g. a UUID. This ‘id’ is unique within a given system, but may or may not be globally unique outside the system. It is used within a system to reference an object from another.
type		string	1..1	The name of the class that is instantiated by a data object representing the Entity.
name		string	0..1	A primary name for the entity.
description		string	0..1	A free-text description of the Entity.
aliases	⋮	string	0..m	Alternative name(s) for the Entity.
extensions	⋮	Extension	0..m	A list of extensions to the Entity, that allow for capture of information not directly supported by elements defined in the model.
subject		Entity \| iriReference	1..1	The Entity or concept about which the Proposition is made.
predicate		string	1..1	The relationship declared to hold between the subject and the object of the Proposition.
object		Entity \| iriReference	1..1	An Entity or concept that is related to the subject of a Proposition via its predicate.

Inherits: Entity

Subclasses: Gene Disease Validity Proposition, Subject Variant Proposition

Used in: Evidence Line, Statement

Implementation Guidance:

1. Structuring Proposition Semantics:

Propositions are abstract representations of possible facts about a domain of discourse, e.g. “NM_005343.4:c.173C>T causes Costello Syndrome”. The semantics of a Proposition are captured in subject, predicate, object, and optional qualifier attributes (SPOQ). Proposition profiles constrain the types of values that can be captured in these SPO attributes, and may define any number of specialized qualifier attributes that extend the SPO “triple” with additional detail or context. For example, if an SPO triple asserts that “NM_005343.4:c.173C>T is causal for Costello Syndrome”, a geneContextQualifier can be used to indicate that its effect on the HRAS gene mediates this causal relationship, and an alleleOriginQualifier can be used to indicate that the fact holds specifically for variants of ‘germline’ origin:

# Note that values in this example are reported in shorthand form for human readability.
# In actual VA-Spec data, many values would be wrapped in complex data type structures such as MappableConcepts.

subject: NM_005343.4:c.173C>T
predicate: isCausalFor
object: Costello Syndrome
geneContextQualifier: HRAS
alleleOriginQualifier: germline

2. Proposition Utility:

A proposition itself makes no claim as to whether the sentiment it expresses is true or not - its job is simply to convey the sharable meaning of a possible fact in a structured data object. Such propositions can then be referenced and reused by Statements and Evidence Lines, which make assertions about them. Specifically:

Statements may report that a proposition was asserted by a particular agent to be true or false, or may report the overall strength of confidence or evidence supporting or disputing a proposition for which a definitive assertion cannot yet be made. Such Statements are based on the agent’s interpretation of evidence as providing discrete argument(s) for or against the proposition.

Evidence Lines represent each such discrete evidence-based argument. An Evidence Line, attached to the Statement it argues about via hasEvidenceLines, reports that a particular collection of information (evidence items) was assessed and scored as evidence to support or dispute some target proposition. It is typically through the assessment of several distinct Evidence Lines that a particular Proposition is ultimately asserted to be true or false in a Statement.

3. Proposition Example

Variant pathogenicity classifications based on the 2015 ACMG Interpretation guidelines are a nice example of how a single proposition such as “NM_005343.4:c.173C>T is causal for Costello Syndrome” can be used in Evidence Lines and Statements. This proposition may first be used as a target against which data is interpreted to build Evidence Lines, according to specific evaluation criteria (e.g. the ACMG PM2 criterion for allele frequency data, and PM1 for functional impact data). As Evidence Lines supporting this proposition accumulate, it may subsequently be used in a Statement where it is asserted as true to classify the variant as ‘pathogenic’.

The example below illustrates how such a scenario may be represented using the VA-Spec - illustrating the use of the same Proposition001 in two Evidence Lines and a Variant Pathogenicity Statement.

# Note that values in this example are reported in shorthand form for human readability.
# In actual VA-Spec data, many values would be wrapped in complex data type structures such as MappableConcepts.

  # As the target proposition of an Evidence Line, based on functional impact data, created at t0 by Curator 1
  id: EvidenceLine001
  type: EvidenceLine
  targetProposition:
    id: VarPathProposition001
    type: VariantPathogenicityProposition
    subject: NM_005343.4:c.173C>T
    predicate: isCausalFor
    object: Costello Syndrome
    geneContextQualifier: HRAS
  hasEvidenceItems: FunctionalImpactStudyResult001       # full StudyResult object omitted for space
  directionOfEvidenceProvided: supports
  strengthOfEvidenceProvided: moderate
  evidenceOutcome: PM1                                   # the ACMG criterion code this evidence line reports (bare code = met at the criterion's default strength)
  specifiedBy: mutational_hot_spot_and_functional_domain_assessment   # the ACMG method type, which constrains the codes 'evidenceOutcome' may use

  # As the target proposition of an Evidence Line, based on cohort allele frequency data, created at t1 by Curator 2
  id: EvidenceLine002
  type: EvidenceLine
  targetProposition: VarPathProposition001    # no need to duplicate an inlined representation, as this Proposition is already defined in the message.
  hasEvidenceItems: AlleleCohortFrequencyStudyResult001   # full StudyResult object omitted for space
  directionOfEvidenceProvided: supports
  strengthOfEvidenceProvided: moderate
  evidenceOutcome: PM2                      # the ACMG criterion code this evidence line reports
  specifiedBy: population_data_assessment   # the ACMG method type, which constrains the codes 'evidenceOutcome' may use

  # As an asserted proposition in a VariantPathogenicityStatement, created at t2 by Curator 3 who puts forth the proposition as true and classifies the variant as 'pathogenic' based on the Evidence Lines above
  id: Statement001
  type: Statement
  proposition: VarPathProposition001          # no need to duplicate an inlined representation, as this Proposition is already defined in the message.
  direction: supports
  strength: definitive
  outcome: pathogenic
  hasEvidenceLines:
    - EvidenceLine001
    - EvidenceLine002
  specifiedBy: 2015 ACMG Variant Interpretation Guidelines