Phylogenetic complexity of the Kimura 3-parameter model (Q1711925)

The part of computational biology that models evolution is called phylogenetics. An important object with crucial role in phylogenetics is a tree model. It consists of a tree, a finite set of states $S$ and a family $\mathcal{M}$ of transition matrices. The most important case is when $S=\{A,C,G,T\}$ where the basis elements correspond to the four nucleobases of DNA. The models for which $\mathcal{M}$ is a proper subspace of matrices reflect symmetries among elements of $S$, where the symmetries are encoded by the action of a finite group $G$ on $S$. \par The Kimura 3-parameter model is one of the most interesting phylogenetic models, which is given by $S=G=\mathbb{Z}_2\times \mathbb{Z}_2$. In this model we have that $S=\{A,C,G,T\}$ and the action of $G$ reflects the pairing between $(A,G)$ and $(C,T)$. \textit{B. Sturmfels} and \textit{S. Sullivant} [J. Comput. Biol. 12, No. 2, 204--228 (2005; Zbl 1391.13058)] conjectured that the ideals of the algebraic varieties associated to the Kimura 3-parameter model are generated by polynomials of degree at most four. In the present article, the authors prove the above conjecture.