Magnitude of type I error when single-locus linkage analysis is maximized over models: a simulation study

Area of Science:

• Genetics
• Statistical Genetics
• Bioinformatics

Background:

• Maximizing LOD scores across multiple genetic models (method of mod scores, MMLS) inflates type I error compared to single-model analysis.
• Modified LOD scores can offer greater power to detect linkage than standard LOD scores under incorrect genetic models.

Purpose of the Study:

• To quantify the actual magnitude of type I error inflation when using modified LOD scores in realistic genetic scenarios.
• To provide guidelines for adjusting LOD score thresholds to maintain desired significance levels in linkage analysis.

Main Methods:

• Simulated datasets with no linkage were generated under various single-locus models (dominant/recessive) with reduced penetrance.
• Linkage analysis was performed by maximizing LOD scores over penetrance, dominance model, or both simultaneously.
• Significance levels and LOD score thresholds were compared against baseline values (one-sided chi2(1)).

Main Results:

• Maximizing over penetrance only approximately doubled significance levels.
• Maximizing over dominance model showed a smaller increase in significance levels.
• Maximizing over both penetrance and dominance increased significance levels two to three times (up to four times) the baseline.
• Recommended adjustments to the LOD score threshold (Z) are approximately 0.30 units for single parameter maximization and 0.60 units for simultaneous maximization.

Conclusions:

• The study quantifies the type I error inflation associated with modified LOD scores in linkage analysis.
• Adjusted LOD score thresholds (up to Z=3.0) are provided to help researchers balance the trade-offs between statistical power and error rates.
• These findings enable informed decisions regarding the use of modified LOD scores in genetic linkage studies.