Affected sib pair tests in inbred populations.

Area of Science:

• Genetics
• Statistical genetics
• Bioinformatics

Background:

• The affected-sib-pair (ASP) method is a key tool for genetic linkage analysis.
• Existing identity-by-state (IBS) test statistics may be sensitive to population inbreeding.
• Modifications are needed to improve the reliability of linkage detection in populations with inbreeding.

Purpose of the Study:

• To modify existing IBS test statistics to accommodate population inbreeding.
• To evaluate the performance of modified IBS tests regarding power and false positive rates.
• To compare the modified tests against original versions under various disease models and inbreeding levels.

Main Methods:

• Modification of two IBS test statistics proposed by Lange (1986a, 1986b).
• Simulation of genetic data under three disease models (recessive, additive, dominant).
• Evaluation of false positive rates and statistical power at controlled significance levels across varying inbreeding coefficients.

Main Results:

• Modified IBS tests maintain nominal false positive rates even with high inbreeding.
• Lange's original tests show inflated false positive rates when inbreeding is substantial.
• Modified z-test power is stable with inbreeding; modified chi-square test power is also stable.
• Lange's chi-square test power increases with inbreeding, exceeding modified test power at high levels.
• All tests show higher power for recessive disease models compared to additive or dominant models.

Conclusions:

• Allowing for inbreeding in IBS tests is crucial for accurate linkage analysis, particularly for recessive diseases.
• Modified IBS tests provide more reliable results in populations with significant inbreeding.
• The choice of test statistic (z-test vs. chi-square) can influence performance depending on inbreeding levels and disease model.