A Model of Compound Heterozygous, Loss-of-Function Alleles Is Broadly Consistent with Observations from...

Area of Science:

• Population Genetics
• Human Complex Disease Genetics
• Quantitative Trait Loci (QTL) Analysis

Background:

• A significant portion of complex disease risk in humans is genetically determined but remains unexplained.
• The allelic spectrum of genetic variants contributing to complex disease risk is largely unknown.
• Theoretical models of population genetics can guide experimental design for studying genetic variation in quantitative traits.

Purpose of the Study:

• To model a genomic region under selection and mutation to understand the genetic architecture of complex traits.
• To evaluate different genetic and demographic models for their impact on identifying disease risk variants.
• To investigate how gene action models influence genetic variance partitioning and statistical test power.

Main Methods:

• Utilized forward simulation to model a genomic region evolving under recurrent deleterious mutation and Gaussian stabilizing selection.
• Compared multiple genetic and demographic models.
• Assessed various methods for identifying genomic regions associated with complex disease risk.

Main Results:

• The model of gene action significantly impacts the population genetic architecture of complex traits, including variance partitioning and statistical test power.
• Models incorporating partial recessivity accurately reflect the minor allele frequency distribution of genome-wide association study (GWAS) hits.
• A gene-based model of incomplete recessivity explains observed discrepancies in heritability estimates and predicts substantial unexplained variance due to intralocus epistasis.

Conclusions:

• Incomplete recessivity provides a plausible explanation for unexplained genetic variance in complex diseases.
• The proposed gene-based recessivity model reconciles empirical observations in heritability studies.
• There is a need for advanced statistical methodologies for region-based genetic association and heritability estimation.