Define more specific structure for inheritance patterns

[caufieldjh]

We could add more specifics to the schema for genetic inheritance patterns.

The class currently has only three slots:

  Inheritance:
    slots:
      - name
      - evidence
      - description

This does not capture details like penetrance, expressivity, polygenicity, or epigenetics, among other factors influencing inheritance and disease states.

Original comment:
[no action required for agent] we should have a more structured way to capture this

Originally posted by @cmungall in #235 (comment)

[github-actions]

Issue Summary: Inheritance Pattern Schema Enhancement

This issue proposes enhancing the Inheritance class to capture more structured information about genetic inheritance patterns beyond the current minimal structure.

Current State

The Inheritance class (lines 2253-2257 in src/dismech/schema/dismech.yaml) currently has only three slots:

Inheritance:
  slots:
    - name
    - evidence
    - description

Real-World Usage Examples

Looking at current disorder files, inheritance information is captured in free text within the description field:

Holt-Oram Syndrome (kb/disorders/Holt_Oram.yaml:296-300):

inheritance:
  - name: Autosomal Dominant
    description: >
      Holt-Oram syndrome is inherited in an autosomal dominant pattern, with affected individuals
      typically being heterozygous carriers of TBX5 mutations. Penetrance is high but variable expressivity
      occurs even within families.

Achondroplasia (kb/disorders/Achondroplasia.yaml:22-27):

inheritance:
  - name: Autosomal dominant
    description: >
      Achondroplasia follows autosomal dominant inheritance with complete penetrance.
      Approximately 80% of cases arise from de novo mutations, typically in the paternal
      germline with advanced paternal age as a risk factor. Homozygous achondroplasia
      is lethal.

These examples mention structured concepts like:

• Penetrance (high, complete, incomplete)
• Expressivity (variable)
• De novo mutations (percentage, parental origin)
• Zygosity effects (heterozygous, homozygous lethal)

Missing Structured Fields

The following inheritance-related concepts are currently only capturable in free text:

1. Penetrance - The proportion of individuals with a pathogenic variant who exhibit the associated phenotype

   • Complete vs incomplete penetrance
   • Age-dependent penetrance
   • Sex-specific penetrance

2. Expressivity - The range of phenotypic manifestations in individuals with the same genotype

   • Variable vs uniform expressivity
   • Factors affecting expressivity

3. De novo mutation rate - Proportion of cases arising from new mutations vs inherited

4. Parent-of-origin effects

   • Paternal age effect
   • Maternal inheritance preference
   • Genomic imprinting

5. Inheritance modifiers

   • Anticipation (expansion of repeat sequences)
   • Mosaicism
   • Digenic/oligogenic inheritance
   • Modifier genes

6. Polygenicity - Multiple genes contributing to phenotype

   • Polygenic risk scores
   • Gene-gene interactions

7. Epigenetic factors

   • Methylation patterns
   • Chromatin modifications affecting expression

Related Issues

This issue is closely related to #126 - Make genetics section in schema more specific, which addresses broader genetics modeling needs. Issue #126 already received detailed analysis (see comment) covering:

• Gene representation with HGNC binding
• Variant specificity
• Association types
• Genotype-phenotype relationships

Ontology Resources

The GENO (Genotype Ontology) provides relevant terms for inheritance modeling:

• GENO includes structured representations of inheritance patterns
• Already partially integrated via ClinicalSignificanceEnum (lines 143-160 in schema)
• Could be extended to cover inheritance mode, penetrance, and expressivity

The Human Phenotype Ontology (HP) also includes inheritance terms under HP:0000005 (Mode of inheritance), such as:

• HP:0000006 - Autosomal dominant inheritance
• HP:0000007 - Autosomal recessive inheritance
• HP:0001417 - X-linked inheritance
• HP:0003743 - Genetic anticipation
• HP:0003745 - Sporadic
• HP:0003829 - Incomplete penetrance
• HP:0003828 - Variable expressivity

Proposed Enhancement Approach

Several options to consider:

1. Add structured slots to Inheritance class

   Inheritance:
     slots:
       - name
       - inheritance_term  # Bind to HP/GENO inheritance mode terms
       - penetrance  # enum: complete, incomplete, variable
       - penetrance_percentage  # numeric for incomplete penetrance
       - expressivity  # enum: variable, uniform
       - de_novo_rate  # percentage
       - parent_of_origin_effect  # free text or structured
       - anticipation  # boolean
       - evidence
       - description

2. Create separate descriptor classes

   • InheritanceModeDescriptor (with HP/GENO term binding)
   • PenetranceDescriptor
   • ExpressivityDescriptor

3. Use qualifiers/modifiers pattern (similar to existing Descriptor class pattern)

Recommendations

1. Start with Option 1 (adding structured slots) as it's the most straightforward and maintains backward compatibility
2. Make ontology term binding optional (recommended: true) to allow gradual migration
3. Bind inheritance_term to HP Mode of Inheritance terms (HP:0000005 descendants)
4. Add quantitative fields for penetrance percentage when not complete
5. Keep free-text description field for nuances not captured by structured fields

Example Enhanced Structure

inheritance:
  - name: Autosomal Dominant
    inheritance_term:
      preferred_term: Autosomal dominant inheritance
      term:
        id: HP:0000006
        label: Autosomal dominant inheritance
    penetrance: INCOMPLETE
    penetrance_percentage: 80-90
    expressivity: VARIABLE
    de_novo_rate: 80
    parent_of_origin_effect: Paternal age effect, typically paternal germline origin
    description: >
      Achondroplasia follows autosomal dominant inheritance with approximately 80-90% penetrance.
      Variable expressivity occurs. Approximately 80% of cases arise from de novo mutations,
      typically in the paternal germline with advanced paternal age as a risk factor.
    evidence:
      - reference: PMID:7913883
        supports: SUPPORT
        snippet: "..."
        explanation: "..."

This enhancement would enable:

• Computational analysis of penetrance patterns across diseases
• Systematic querying of de novo mutation rates
• Better representation of complex inheritance patterns
• Improved interoperability with genetic databases

Would this approach align with the project's goals for structured genetic information capture?