Inheritance Pattern Analysis

Variant-Linker provides powerful inheritance pattern analysis capabilities, helping researchers identify variants that follow expected inheritance patterns in family studies.

Overview

The inheritance analysis module automatically examines genotype data and family relationships to deduce potential inheritance patterns for variants. This feature is essential for:

• Clinical Genetics: Identifying causative variants in family studies
• Research: Filtering variants based on inheritance expectations
• Variant Prioritization: Ranking variants by inheritance pattern compatibility
• Quality Control: Validating family relationships and genotype data

Supported Inheritance Patterns

De Novo Variants

Variants present in the child but absent in both parents:

• Indicates potential new mutations
• High priority for disease causation studies
• Requires trio or extended family data

Autosomal Dominant (AD)

Heterozygous variants that segregate with affected status:

• One copy of variant causes phenotype
• Affected individuals typically have one affected parent
• 50% transmission rate to offspring

Autosomal Recessive (AR)

Homozygous variants in affected individuals with carrier parents:

• Two copies required for phenotype
• Parents typically unaffected carriers
• 25% transmission rate for affected offspring from carrier parents

X-linked Dominant (XLD)

Variants on X chromosome following dominant inheritance:

• Affects both males and females
• Affected males pass trait to all daughters
• Affected females have 50% transmission rate

X-linked Recessive (XLR)

Variants on X chromosome following recessive inheritance:

• Primarily affects males
• Carrier mothers transmit to 50% of sons
• Affected fathers pass carrier status to all daughters

Compound Heterozygous

Two different variants in the same gene that together cause recessive phenotype:

• Each variant inherited from different parents
• Gene-level analysis required
• Both variants must affect gene function

Analysis Modes

Single Sample Mode

When only one sample is present in the VCF file:

• Limited inheritance pattern analysis
• Focuses on variant annotation and basic pattern possibilities
• Cannot determine de novo status or segregation

Trio Mode

When parent-child trio is available:

• Full de novo analysis
• Basic dominant/recessive pattern assessment
• Optimal for most clinical applications

Trio specification methods:

# Using PED file
variant-linker --vcf-input trio.vcf --ped trio.ped --calculate-inheritance

# Manual specification (Index, Mother, Father)
variant-linker --vcf-input trio.vcf --sample-map "PROBAND,MOTHER,FATHER" --calculate-inheritance

Extended Family Mode

When comprehensive family structure is provided via PED file:

• Multi-generational inheritance analysis
• Complex pattern recognition
• Compound heterozygous detection
• Inheritance pattern validation across generations

# Extended family analysis
variant-linker --vcf-input extended_family.vcf --ped extended_family.ped --calculate-inheritance

Usage Examples

Basic Trio Analysis

# Simple trio analysis
variant-linker \
  --vcf-input family_trio.vcf \
  --ped trio.ped \
  --calculate-inheritance \
  --output JSON \
  --save trio_results.json

Extended Family Study

# Multi-generational family analysis
variant-linker \
  --vcf-input large_family.vcf \
  --ped large_family.ped \
  --calculate-inheritance \
  --output VCF \
  --save annotated_with_inheritance.vcf

Compound Heterozygous Analysis

# Gene-level compound heterozygous detection
variant-linker \
  --vcf-input family.vcf \
  --ped family.ped \
  --calculate-inheritance \
  --vep_params "CADD=1,SIFT=1,PolyPhen=1" \
  --output CSV \
  --save compound_het_analysis.csv

Output Format

Inheritance Analysis Results

When inheritance analysis is enabled, each variant annotation includes a deducedInheritancePattern object:

{
  "deducedInheritancePattern": {
    "patterns": ["autosomal_recessive", "compound_heterozygous"],
    "confidence": "high",
    "patternDetails": {
      "segregationStatus": "consistent",
      "affectedCarriers": 2,
      "unaffectedCarriers": 0,
      "genotypeCounts": {
        "homozygous_ref": 2,
        "heterozygous": 4,
        "homozygous_alt": 1
      }
    }
  }
}

Pattern Fields

• patterns: Array of possible inheritance patterns
• confidence: Confidence level (high, medium, low)
• patternDetails: Additional analysis information
• segregationStatus: Whether variant segregates with phenotype
• genotypeCounts: Genotype distribution in family

Confidence Levels

High Confidence

• Clear segregation pattern
• Sufficient family members
• Consistent with single inheritance mode

Medium Confidence

• Some evidence for pattern
• Limited family size or incomplete data
• Multiple possible patterns

Low Confidence

• Insufficient data for pattern determination
• Conflicting evidence
• Complex inheritance not clearly resolved

Algorithm Details

Pattern Detection Logic

1. Genotype Extraction: Extract genotypes for all family members
2. Relationship Mapping: Map family relationships from PED data
3. Segregation Analysis: Analyze variant transmission patterns
4. Pattern Matching: Compare observed patterns to expected inheritance models
5. Confidence Assessment: Evaluate strength of evidence for each pattern

Core Calculation Algorithm

The inheritance analysis engine uses a multi-step calculation process:

1. Index Sample Determination

The system identifies the proband/index sample using this priority:

• Explicit sampleMap with 'index' or 'proband' designation
• First affected individual (phenotype = '2') in PED file
• First sample ID in the VCF genotype data

2. Genotype Classification

Each genotype is classified based on allele counts:

• 0/0 or 0|0: Homozygous reference (isRef)
• 0/1 or 1/0 or 0|1 or 1|0: Heterozygous (isHet)
• 1/1 or 1|1: Homozygous alternate (isHomAlt)
• ./. or .: Missing genotype
• Multi-allelic calls handled by checking specific allele indices

3. Pattern-Specific Rules

De Novo Detection:

IF index has variant (Het or HomAlt) AND
   mother is HomRef (0/0) AND
   father is HomRef (0/0)
THEN pattern = de_novo

Autosomal Dominant (AD):

IF index is Het AND
   (mother is Het OR father is Het) AND
   NOT both parents are Het
THEN pattern = autosomal_dominant

Autosomal Recessive (AR):

IF index is HomAlt AND
   mother is Het AND
   father is Het
THEN pattern = autosomal_recessive

X-linked Patterns (requires chromosome information):

• X-linked Dominant (XLD):
  • Affected males: Must have affected mother
  • Affected females: One affected parent sufficient
• X-linked Recessive (XLR):
  • Affected males: Carrier mother (Het) or affected mother
  • Affected females: Both X chromosomes must carry variant

4. Segregation Scoring

For extended families, segregation consistency is calculated:

segregation_score = (carriers_affected + non_carriers_unaffected) / total_informative_members

Where:

• carriers_affected: Variant carriers who are affected
• non_carriers_unaffected: Non-carriers who are unaffected
• total_informative_members: All family members with both genotype and phenotype data

Segregation categories:

• Perfect (score = 1.0): Complete segregation with phenotype
• High (score ≥ 0.8): Strong segregation evidence
• Moderate (score ≥ 0.6): Some segregation support
• Low (score < 0.6): Poor segregation

5. Confidence Calculation

Confidence levels integrate multiple factors:

confidence = calculateConfidence({
  segregationScore,
  familySize,
  missingDataCount,
  mendelianErrors,
  patternConsistency,
});

Confidence modifiers:

• High: Large families (≥5), perfect segregation, no missing data
• Medium: Small families (3-4), good segregation (≥0.8), minimal missing data
• Low: Duo analysis, poor segregation (<0.8), significant missing data

De Novo Detection

De novo variants are identified when:

• Child has variant (heterozygous or homozygous)
• Both parents lack the variant (homozygous reference)
• High-quality genotype calls for all family members
• No evidence of sample mix-up or technical errors

Additional checks for de novo calls:

• Allele balance: Het calls should have ~50% alternate allele frequency
• Read depth: Sufficient coverage in all trio members
• Quality scores: High genotype quality (GQ) values

Compound Heterozygous Detection

Compound heterozygous variants are identified by:

1. Gene-level Analysis: Group variants by affected gene
2. Phase Analysis: Determine if variants are on different chromosomes
3. Parent-of-Origin: Verify variants inherited from different parents
4. Functional Impact: Both variants must potentially affect gene function

Algorithm for compound het detection:

FOR each gene with multiple Het variants in index:
  IF variant1 from mother (mother Het, father HomRef) AND
     variant2 from father (father Het, mother HomRef) AND
     both variants have functional impact
  THEN compound_heterozygous = true

Multi-Allelic Variant Handling

For variants with multiple alternate alleles:

1. Split multi-allelic sites into biallelic representations
2. Analyze each alternate allele independently
3. Preserve allele-specific genotype information
4. Report inheritance patterns per alternate allele

Example:

Original: 1:12345 A T,G (GT: 1/2)
Split to:
  - 1:12345 A T (GT: 1/0)
  - 1:12345 A G (GT: 0/1)

Special Cases and Edge Conditions

Hemizygous Calls (Male X Chromosome)

• Males have single X chromosome
• Genotypes reported as "1" (not "1/1")
• Special handling for X-linked inheritance calculations

Missing Genotypes

• Pattern reported as "unknown_missing_genotype"
• Segregation analysis excludes missing samples
• Confidence automatically reduced

Technical Replicates

• If duplicate samples detected, use highest quality call
• Flag inconsistent genotypes between replicates

Quality Control

The analysis includes several quality control measures:

• Genotype Quality Filtering: Remove low-quality genotype calls
• Mendelian Error Detection: Identify inconsistent inheritance
• Sample Relationship Validation: Verify expected family relationships
• Technical Artifact Filtering: Remove likely technical errors

Quality metrics tracked:

• Mendelian error rate: Percentage of impossible inheritance patterns
• Missing data rate: Proportion of missing genotypes
• Hardy-Weinberg equilibrium: For population-level checks
• Allele balance: For heterozygous call validation

Limitations and Considerations

Technical Limitations

• Incomplete Penetrance: Not explicitly modeled in current version
• Genomic Imprinting: Parent-of-origin effects not fully considered
• Structural Variants: Limited support for complex structural variations
• Mosaic Variants: Low-level mosaicism may not be detected

Data Requirements

• High-Quality Genotypes: Reliable genotype calls essential
• Complete Pedigree: Missing family members reduce analysis power
• Accurate Phenotyping: Affected status must be correctly assigned
• Consistent Sample Naming: Sample IDs must match between VCF and PED files

Interpretation Guidelines

1. Multiple Patterns: Variants may show evidence for multiple inheritance patterns
2. Confidence Levels: Consider confidence when interpreting results
3. Functional Validation: Inheritance pattern alone insufficient for causality
4. Clinical Context: Integrate with clinical and functional evidence

Implementation Architecture

Module Organization

The inheritance analysis system is composed of specialized modules:

Core Modules

• inheritanceAnalyzer.js: Main orchestrator that coordinates all analysis
• patternDeducer.js: Deduces potential inheritance patterns from genotypes
• segregationChecker.js: Validates pattern consistency across families
• patternPrioritizer.js: Ranks patterns by likelihood and evidence
• compoundHetAnalyzer.js: Detects compound heterozygous variants

Utility Modules

• genotypeUtils.js: Genotype parsing and classification functions
• pedigreeUtils.js: Family relationship and sex determination utilities

Data Flow

1. VCF + PED Input
   ↓
2. Genotype Extraction (per variant)
   ↓
3. Pattern Deduction
   - Single sample analysis
   - Trio analysis
   - Extended family analysis
   ↓
4. Segregation Checking
   - Calculate segregation scores
   - Validate Mendelian inheritance
   ↓
5. Pattern Prioritization
   - Rank by evidence strength
   - Apply confidence scores
   ↓
6. Compound Het Analysis
   - Group by gene
   - Check parent-of-origin
   ↓
7. Results Integration
   - Merge all analyses
   - Format output

Key Data Structures

Genotype Map

Map<variantKey, Map<sampleId, genotype>>
// Example:
// "1-12345-A-T" → { "SAMPLE1": "0/1", "SAMPLE2": "1/1", "SAMPLE3": "0/0" }

Pedigree Data

Map<sampleId, {
  familyId: string,
  paternalId: string,
  maternalId: string,
  sex: string,        // "1" (male), "2" (female), "0" (unknown)
  phenotype: string   // "1" (unaffected), "2" (affected), "0" (unknown)
}>

Pattern Result

{
  patterns: ["autosomal_recessive", "compound_heterozygous"],
  confidence: "high",
  segregation: {
    score: 0.95,
    category: "high",
    details: { /* segregation metrics */ }
  },
  priorityScore: 0.85
}

Advanced Features

Custom Pattern Definitions

Future versions will support custom inheritance pattern definitions for:

• Disease-specific inheritance models
• Population-specific patterns
• Complex multi-gene interactions

Integration with Functional Annotation

Inheritance analysis integrates with functional annotation to prioritize variants:

• Functional Impact: Consider variant consequence severity
• Gene Constraint: Integrate gene constraint metrics
• Pathogenicity Scores: Weight by computational pathogenicity predictions

Performance Optimization

For large-scale analysis:

• Parallel Processing: Variants analyzed concurrently
• Memory Efficiency: Streaming VCF processing
• Caching: Pedigree relationships cached for reuse
• Early Termination: Skip analysis for obvious non-pathogenic variants

Troubleshooting

Common Issues

No Inheritance Patterns Detected

• Check sample naming consistency between VCF and PED files
• Verify family relationships in PED file
• Ensure genotype quality is sufficient

Inconsistent Patterns

• Review phenotype assignments in PED file
• Check for sample mix-ups or labeling errors
• Consider incomplete penetrance or variable expressivity

Low Confidence Results

• Increase family size if possible
• Improve genotype quality through better sequencing
• Validate relationships with independent methods

Debug Mode

Use debug mode to troubleshoot inheritance analysis:

variant-linker \
  --vcf-input family.vcf \
  --ped family.ped \
  --calculate-inheritance \
  --debug 3 \
  --output JSON

Debug output includes:

• Genotype extraction details
• Family relationship parsing
• Pattern matching logic
• Confidence calculation steps

Best Practices

Study Design

1. Family Selection: Choose informative family structures
2. Sample Quality: Ensure high-quality DNA and sequencing
3. Phenotype Definition: Use consistent, well-defined phenotypes
4. Control Samples: Include unaffected family members when possible

Data Processing

1. Genotype Filtering: Apply appropriate quality filters
2. Variant Normalization: Ensure consistent variant representation
3. Reference Consistency: Use consistent reference genome versions
4. Sample Validation: Verify sample identity and relationships

Result Interpretation

1. Multiple Lines of Evidence: Combine inheritance with functional data
2. Population Frequencies: Consider variant frequencies in relevant populations
3. Clinical Context: Integrate with clinical presentation and history
4. Functional Validation: Confirm causality through functional studies

Next Steps

• Explore custom scoring configuration to weight inheritance patterns
• Learn about VCF and PED file preparation
• Check the contributing guidelines for development guidelines