Dissimilarity maps on trees and the representation theory of SL_m( C)

DOI10.1007/S10801-010-0241-9MaRDI QIDQ626435zbMATH OpenOpenAlexFDO

Publication date 18 February 2011

Published in Journal of Algebraic Combinatorics (Search for Journal in Brave)

Copyright license Creative Commons Attribution-NonCommercial 4.0 International

Full work available at URL https://arxiv.org/abs/1003.3970

Trees (05C05) Combinatorial aspects of representation theory (05E10) Representation theory for linear algebraic groups (20G05)

Abstract: We give another proof that

m

-dissimilarity vectors of weighted trees are points on the tropical Grassmanian, as conjectured by Cools, and proved by Giraldo in response to a question of Sturmfels and Pachter. We accomplish this by relating

m

-dissimilarity vectors to the representation theory of

S L_{m} .

Review text

The main result of this article is that the higher dissimilarity vectors of weighted trees are points on tropical Grassmannians. The meaning of this is as follows. Given an edge-weighted trivalent tree, one can define a metric on the leaves by declaring the distance between two leaves to be the sum of the edge-weights on the unique path connecting them. One can consider all pairs of vectors and place these distances in a vector, the ``2-dissimilarity vector. It is well-known that one can recover the tree from this vector, and fundamental work of \textit{D. Speyer} and \textit{B. Sturmfels} in tropical geometry [Adv. Geom. 4, No. 3, 389--411 (2004; Zbl 1065.14071); \url{arXiv:math/0304218}] showed that the space of 2-dissimilarity vectors actually coincides with the tropical Grassmannian obtained by tropicalizing the Plucker equations defining Gr\((2,n)\), where \(n\) is the number of leaves. Given \(m\geq 2\), one can assign to each \(m\)-tuple of leaves the sum of weights of edges that are part of a path connecting a pair of these leaves. This leads to \(m\)-dissimilarity vectors, and the result of the present paper is that these vectors define a subset of the tropical Grassmannian Gr\((m,n)\). The history of this result is interesting. The question first arose in work of Patcher and Sturmfels in computational biology. Their interest in that context was of a practical matter, not simply theoretical. The work of Speyer and Sturmfels addressed the case \(m=2\), but \(m > 2\) remained open. \textit{C. Bocci} and \textit{F. Cools} [Appl. Math. Comput. 212, No. 2, 349--356 (2009; Zbl 1220.05016); \url{arXiv:0803.2184}] put the higher dissimilarity vectors in a tropical framework, and \textit{F. Cools} [J. Symb. Comput. 44, No. 8, 1079--1086 (2009; Zbl 1207.14061); \url{arXiv:0903.2010}] proved the result for small \(m\) and conjectured it for all \(m\). Subsequently, the general case was proven by \textit{B. I. Giraldo} [Electron. J. Comb. 17, No. 1, Research Paper N6, 7 p. (2010; Zbl 1184.05023); \url{arXiv:0909.0141}] using very combinatorial methods, and several months later the present article was posted on the arXiv, proving the same result with quite different tools. The approach of the present article is to show that in some generality filtrations of branching algebras arising in representation theory can be used to define points on certain tropical varieties. This is based largely on previous work of the author [``The algebra of conformal blocks, preprint, \url{arXiv:0910.0577}] and after developing the necessary tools and results in the first half of the present article, is applied in a special situation to derive the desired result about dissimilarity vectors and the tropical Grassmannian.

Recommendations

Cites work

Cited in

(6)

This page was built for publication: Dissimilarity maps on trees and the representation theory of \(\mathrm{SL}_{m}(\mathbb C)\)

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q626435)