Guess LOD approach: sufficient conditions for robustness

Area of Science:

• Genetics
• Statistical genetics
• Bioinformatics

Background:

• Genetic linkage analysis requires precise genetic models for inheritance and gene frequencies.
• Complex diseases often involve misspecified genetic models, posing challenges for linkage studies.
• Previous work indicated disease model misspecification doesn't inflate false positives if marker alleles are known.

Purpose of the Study:

• To extend previous findings to scenarios with incorrect marker locus genetic models.
• To establish conditions for unbiased recombination fraction estimation and valid statistical tests under linkage null hypothesis.
• To investigate the impact of model misspecification on genetic linkage analysis under various ascertainment schemes.

Main Methods:

• Analytic studies and simulation studies were employed.
• Investigation of genetic linkage under different selection or ascertainment schemes.
• Evaluation of conditions for asymptotic unbiased estimation and chi-squared distribution of the likelihood ratio test.

Main Results:

• Misspecification of the marker locus genetic model can lead to asymptotic bias in recombination fraction estimation.
• Failure to accurately estimate marker allele frequencies is a primary cause of excess false-positive genetic linkage evidence.
• In most scenarios, misspecification of the disease or trait inheritance model alone does not solely generate false-positive linkage results.

Conclusions:

• Accurate marker allele frequency estimation is critical for robust genetic linkage analysis.
• While disease model misspecification is less problematic, marker model misspecification, particularly allele frequencies, can compromise study validity.
• The findings provide crucial insights for refining genetic linkage analysis methodologies, especially for complex genetic traits.