Near-medians that avoid the corners; a combinatorial probability approach.

Area of Science:

• Computational Biology
• Genomics
• Phylogenetics

Background:

• Breakpoint medians in random genomes tend towards individual genomes ('corners') with increasing length.
• A small proportion of medians are equidistant from all genomes ('medians in the middle'), often missed by algorithms.
• This 'corner effect' can bias median calculations in real genomic data with shared adjacencies.

Purpose of the Study:

• To develop a method for constructing a "near median" genome that approximates the true median.
• To address the limitations of existing algorithms in identifying equidistant medians.
• To generalize median approximation for genomes with varying shared gene adjacencies.

Main Methods:

• A simple sampling procedure is proposed to generate a near median genome.
• The method calculates breakpoint distances to approximate median status.
• A supplementary sampling scheme is introduced to further refine the median approximation.

Main Results:

• The proposed sampling procedure yields a near median genome with a breakpoint distance only slightly larger than the true median.
• The method is generalized to handle genomes with shared gene adjacencies.
• A supplementary scheme further improves the proximity to median status.

Conclusions:

• The developed approach is valuable for phylogenetic analyses requiring repeated median calculations at ancestral nodes.
• It mitigates the 'corner effect' that can hinder communication between different parts of a phylogenetic tree.
• This method offers a practical solution for inferring ancestral genomic states.