Compound Poisson approximation and testing for gene clusters with multigene families

Area of Science:

• Computational Biology
• Genomics
• Statistical Genetics

Background:

• Comparative genomics relies on identifying conserved genomic regions across species.
• Significance testing for these regions is crucial for evolutionary and functional insights.
• Existing methods may not fully account for complex genomic structures like multigene families.

Purpose of the Study:

• To develop and present a compound Poisson approximation for significance testing of conserved genomic regions.
• To incorporate the influence of multigene families into these statistical computations.
• To provide convergence results for the approximation's error.

Main Methods:

• Utilizing a compound Poisson approximation framework.
• Applying the Stein-Chen method to derive error bounds for the approximation.
• Focusing on the reference region approach for identifying conserved segments.

Main Results:

• A novel compound Poisson approximation is established for significance testing in comparative genomics.
• Convergence results for the approximation's error are obtained, validating its accuracy.
• The method explicitly addresses the challenge of multigene families in conserved region detection.

Conclusions:

• The proposed compound Poisson approximation offers a robust computational tool for significance testing of conserved genomic regions.
• Accounting for multigene families enhances the precision of these tests.
• The Stein-Chen method provides a rigorous basis for the approximation's error analysis.