Haplotype uncertainty in association studies

Area of Science:

• Population Genetics
• Genetic Epidemiology
• Bioinformatics

Background:

• Haplotype inference from genotype data is fundamental in population genetic association studies.
• Acknowledging and addressing uncertainty in haplotype inference is recognized as critical for study validity.

Purpose of the Study:

• To evaluate the effectiveness of simple correction methods for haplotype inference uncertainty.
• To identify genetic region characteristics susceptible to haplotype uncertainty influences.
• To assess bias mitigation strategies in haplotype inference, particularly in case-control studies.

Main Methods:

• Utilized PHASE methodology for haplotype inference.
• Conducted case-control association analyses for non-Hodgkin lymphoma.
• Employed simulations to model varying degrees of haplotype uncertainty, linkage, and missing genotype data.

Main Results:

• Minimal impact of uncertainty adjustment was observed in non-Hodgkin lymphoma association analyses.
• Bias in haplotype inference can be avoided by using haplotype probabilities or multiple imputation.
• Separate inference for case and control populations is essential for these bias mitigation methods.

Conclusions:

• While simple adjustments had limited effect, robust methods like haplotype probabilities and multiple imputation can effectively manage haplotype uncertainty.
• Multiple imputation offers the advantage of incorporating uncertainty into confidence interval estimations.
• Findings provide guidance for handling haplotype inference complexity in large, marker-rich genetic regions.