The promise of composite likelihood methods for addressing computationally intensive challenges

Area of Science:

• Genetics
• Statistical Inference
• Computational Biology

Background:

• High-dimensional genetic data presents complex correlation structures, challenging traditional statistical inference methods.
• Standard likelihood-based approaches struggle with the complexity of genetic datasets.
• Composite likelihood offers a viable approximation strategy for genetic data analysis.

Purpose of the Study:

• To describe composite likelihood methods for statistical inference in genetics.
• To review current applications of composite likelihood in genetic studies.
• To highlight future research directions for composite likelihood in genetics.

Main Methods:

• Description of composite likelihood methodology.
• Review of existing genetic applications.
• Discussion of inferential advantages and challenges.

Main Results:

• Composite likelihood is a successful approximation for certain genetic applications.
• The method shows potential for broader application in genetic research.
• Identified key areas for future research in composite likelihood for genetics.

Conclusions:

• Composite likelihood is a valuable tool for statistical inference with high-dimensional genetic data.
• Further research is essential to fully leverage composite likelihood in population genetics and disease gene mapping.
• The method facilitates estimation of parameters like recombination rate and linkage disequilibrium.