[Paper Review] A note on convex characters and Fibonacci numbers

Steven Kelk|arXiv (Cornell University)|Aug 11, 2015

Genomics and Phylogenetic Studies9 references3 citations

TL;DR

This paper establishes that the number of convex characters on any unrooted binary phylogenetic tree with n taxa is the (2n−1)th Fibonacci number, independent of tree topology. It further shows that the number of convex characters with each state on at least two taxa is the (n−1)th Fibonacci number, enabling a faster algorithm for the maximum parsimony distance problem with time complexity Θ(φⁿ · poly(n)), where φ is the golden ratio.

ABSTRACT

Given an unrooted, binary phylogenetic tree T on a set of n >= 2 taxa, a closed expression for the number of convex characters on T has been known since 1992, and this is independent of the exact topology of T. Here we prove that this number is equal to the (2n-1)th Fibonacci number. Moreover, we show that the number of convex characters in which each state appears on at least two taxa, is also independent of topology, and equal to the (n-1)th Fibonacci number. We use this insight to give a simple but effective algorithm for the NP-hard maximum parsimony distance problem that runs in time $\Theta( \phi^{n} \cdot ext{poly}(n) )$, where $\phi \approx 1.618...$ is the golden ratio. Finally, we give an explicit example demonstrating that topological neutrality no longer holds when counting the number of convex characters in which each state appears on at least three taxa.

Motivation & Objective

• To determine whether the number of convex characters on unrooted binary trees depends on tree topology.
• To establish a closed-form expression for the number of convex characters in terms of Fibonacci numbers.
• To develop an efficient algorithm for the NP-hard maximum parsimony distance problem using Fibonacci-based enumeration.
• To investigate whether topological neutrality persists when requiring each character state to appear on at least three taxa.

Proposed method

• Proves that the number of convex characters on any unrooted binary tree with n taxa is the (2n−1)th Fibonacci number, regardless of topology.
• Demonstrates that the number of convex characters with each state on at least two taxa is the (n−1)th Fibonacci number, again independent of topology.
• Uses the Fibonacci structure to design a dynamic programming algorithm for maximum parsimony distance with time complexity Θ(φⁿ · poly(n)).
• Applies the golden ratio φ ≈ 1.618 to bound the exponential growth of the search space.
• Employs recursive decomposition of character states and tree structure to derive the Fibonacci identities.
• Constructs a concrete counterexample to show that topological neutrality fails when each state must appear on at least three taxa.

Experimental results

Research questions

• RQ1 Is the number of convex characters on an unrooted binary tree independent of its topology?
• RQ2 Can the number of convex characters with each state on at least two taxa be expressed as a Fibonacci number?
• RQ3 Does the topological neutrality of convex character counts persist when requiring each state to appear on at least three taxa?
• RQ4 Can Fibonacci-based enumeration be leveraged to design faster algorithms for phylogenetic distance problems?
• RQ5 What is the exact combinatorial structure of convex characters under state frequency constraints?

Key findings

• The number of convex characters on any unrooted binary tree with n taxa is exactly the (2n−1)th Fibonacci number, regardless of tree topology.
• The number of convex characters in which each state appears on at least two taxa is the (n−1)th Fibonacci number, and this count is also topologically invariant.
• An algorithm for the maximum parsimony distance problem runs in time Θ(φⁿ · poly(n)), where φ ≈ 1.618 is the golden ratio.
• The exponential factor φⁿ is optimal in the sense that it matches the growth rate of Fibonacci numbers.
• A counterexample is constructed showing that topological neutrality breaks down when each state must appear on at least three taxa.
• The Fibonacci-based characterization provides a tight combinatorial bound for convex character enumeration under various constraints.