MFLINK

MFLINK performs likelihood-based, nearly model-free linkage analysis by maximizing lod score likelihoods over transmission and linkage parameters to detect genetic effects while accommodating transmission model misspecification.

Key Features:

• Model-Free Approach: Implements a nearly model-free calculation of lod scores by maximizing likelihoods over both transmission and linkage parameters to mitigate bias from transmission model misspecification, particularly at small recombination fractions and in multipoint analyses.
• Transmission Model Flexibility: Independently maximizes likelihoods under various transmission models, including Mendelian dominant, null effect, and Mendelian recessive models, to avoid constraints from specific model assumptions.
• Population Prevalence Constraint: Constrains transmission model parameters to produce the correct population prevalence to prevent biased parameter estimates and preserve test power.
• Admixture Maximization: Maximizes the likelihood under linkage over an admixture parameter representing the proportion of families linked at a given map position.
• Versatility Across Data Types: Applies to affected sib pairs and pedigrees and yields robust lod scores at test positions close to markers in multipoint analyses.

Scientific Applications:

• Linkage Detection: Detects linkage when a genetic effect is present and evaluates evidence under the null hypothesis.
• Genetic Research: Supports analysis of diverse inheritance scenarios and family structures, including complex inheritance patterns.

Methodology:

Maximizes likelihoods over a limited set of transmission models while independently considering linkage parameters; maximizes linkage likelihood over an admixture parameter representing the proportion of linked families; constrains transmission parameters to produce the correct population prevalence and derives lod scores from the resulting likelihoods.