The heterogeneity problem. I: Separating genetic from environmental forms of the same disease

Area of Science:

• Quantitative genetics
• Statistical genetics
• Genetic epidemiology

Background:

• Segregation analysis is crucial for understanding disease inheritance patterns.
• Detecting heterogeneity, where a disease has multiple causes (genetic and environmental), is challenging.
• Existing methods may struggle with accurately quantifying the proportion of genetic cases in mixed datasets.

Purpose of the Study:

• To reparameterize the likelihood function for segregation analysis to better detect heterogeneity.
• To develop a parameter (alpha) estimating the proportion of genetic disease families within a dataset.
• To validate the reparameterized method using simulations under various genetic and environmental models.

Main Methods:

• Developed a novel likelihood reparameterization for segregation analysis.
• Simulated nuclear family data with both Mendelian recessive genetic and random environmental disease models.
• Estimated the proportion of genetic families (alpha) and compared it with underlying genetic parameters (gene frequency q, sporadic frequency R).

Main Results:

• The reparameterized method accurately estimated the proportion of genetic families (alpha), with deviations of only a few percent from true values.
• Estimates for underlying genetic parameters (gene frequency q, sporadic frequency R) were less reliable, ranging from fair to poor.
• Segregation analysis effectively estimated disease heterogeneity even with unreliable estimates of gene frequency and sporadic penetrance.

Conclusions:

• The reparameterized segregation analysis provides a robust method for quantifying disease heterogeneity.
• This approach accurately determines the proportion of genetic disease families in mixed etiology datasets.
• The method is valuable for genetic epidemiology research, offering reliable heterogeneity estimates independent of precise genetic parameter accuracy.