Inherent intractability of the ascertainment problem for pedigree data: a general likelihood framework

Area of Science:

• Genetics
• Statistical Genetics
• Bioinformatics

Background:

• Ascertainment bias occurs when families are selected based on affected individuals.
• Correcting for ascertainment is crucial but lacks methods for many sampling schemes, like sequential pedigree sampling.
• Investigating the necessity of ascertainment correction for large pedigrees and multiple probands is ongoing.

Purpose of the Study:

• To address challenges in ascertainment correction for segregation analysis.
• To introduce and analyze proband-independent (PI) and proband-dependent (PD) sampling.
• To evaluate the impact of pedigree structure and size on ascertainment correction.

Main Methods:

• Distinguishing between PI and PD sampling based on pedigree structure dependency.
• Analyzing the validity of conditioning on observed pedigree structure under different sampling schemes.
• Examining the feasibility of PI sampling for large pedigrees.

Main Results:

• Conditioning on pedigree structure is generally incorrect for PD sampling, except for single selection.
• PI sampling of large pedigrees is often not feasible.
• Current methods for genetic modeling in ascertained data are frequently intractable.

Conclusions:

• The problem of ascertainment correction in genetic analysis is often intractable with current methods.
• Standard approaches fail for complex pedigrees and sequential sampling strategies.
• Development of robust approximate methods for ascertainment correction is recommended.