Search

UPGMA (Unweighted Pair Group Method with Arithmetic Mean) is a widely used clustering method. Here we show that UPGMA is a greedy heuristic for the normalized equidistant minimum evolution (NEME) problem, that is, finding a rooted tree that minimizes the minimum evolution score relative to the dissimilarity matrix among all rooted trees with the same leaf-set in which all leaves have the same distance to the root. We prove that the NEME problem is NP-hard. In addition, we present some heuristic and approximation algorithms for solving the NEME problem, including a polynomial time algorithm that yields a binary, rooted tree whose NEME score is within O(log^2 n) of the optimum. We expect that these results to eventually provide further insights into the behavior of the UPGMA algorithm., Comment: 29 pages, 8 figures

Tandem duplication is a rearrangement process whereby a segment of DNA is replicated and proximally inserted. A sequence of these events is termed an evolution. Many different configurations can arise from such evolutions, generating some interesting combinatorial properties. Firstly, new DNA connections arising in an evolution can be algebraically represented with a word producing automaton. The number of words arising from n tandem duplications can then be recursively derived. Secondly, many distinct evolutions result in the same sequence of words. With the aid of a bi-colored 2d-tree, a Hasse diagram corresponding to a partially ordered set is constructed, for which the number of linear extensions equates to the number of evolutions generating a given word sequence. Thirdly, we implement some subtree prune and graft operations on this structure to show that the total number of possible evolutions arising from n tandem duplications is ? k = 1 n ( 4 k - ( 2 k + 1 ) ) . The space of structures arising from tandem duplication thus grows at a super-exponential rate with leading order term O ( 4 1 2 n 2 ) .

In phylogenetics, a common strategy used to construct an evolutionary tree for a set of species $$X$$ is to search in the space of all such trees for one that optimizes some given score function (such as the minimum evolution, parsimony or likelihood score). As this can be computationally intensive, it was recently proposed to restrict such searches to the set of all those trees that are compatible with some circular ordering of the set $$X$$ . To inform the design of efficient algorithms to perform such searches, it is therefore of interest to find bounds for the number of trees compatible with a fixed ordering in the neighborhood of a tree that is determined by certain tree operations commonly used to search for trees: the nearest neighbor interchange (nni), the subtree prune and regraft (spr) and the tree bisection and reconnection (tbr) operations. We show that the size of such a neighborhood of a binary tree associated with the nni operation is independent of the tree’s topology, but that this is not the case for the spr and tbr operations. We also give tight upper and lower bounds for the size of the neighborhood of a binary tree for the spr and tbr operations and characterize those trees for which these bounds are attained.

To any metric D on a finite set X, one can associate a metric space T(D) known as its tight span. Properties of T(D) often reveal salient properties of D. For example, cut sets of T(D), i.e., subsets of T(D) whose removal disconnect T(D), can help to identify clusters suggested by D and indicate how T(D) (and hence D) may be decomposed into simpler components. Given a bipartition or split S of X, we introduce in this paper a real-valued index @e"("D","S") that comes about by considering cut sets of T(D). We also show that this index is intimately related to another, more easily computable index @d"("D","S") whose definition does not directly depend on T(D). In addition, we provide an illustration for how these two new indices could help to extend and complement current distance-based methods for phylogenetic network construction such as split decomposition and NeighborNet.

Phylogenetic networks are a generalization of evolutionary trees that are used by biologists to represent the evolution of organisms which have undergone reticulate evolution. Essentially, a phylogenetic network is a directed acyclic graph having a unique root in which the leaves are labelled by a given set of species. Recently, some approaches have been developed to construct phylogenetic networks from collections of networks on 2- and 3-leaved networks, which are known as binets and trinets, respectively. Here we study in more depth properties of collections of binets, one of the simplest possible types of networks into which a phylogenetic network can be decomposed. More specifically, we show that if a collection of level-1 binets is compatible with some binary network, then it is also compatible with a binary level-1 network. Our proofs are based on useful structural results concerning lowest stable ancestors in networks. In addition, we show that, although the binets do not determine the topology of the network, they do determine the number of reticulations in the network, which is one of its most important parameters. We also consider algorithmic questions concerning binets. We show that deciding whether an arbitrary set of binets is compatible with some network is at least as hard as the well-known graph isomorphism problem. However, if we restrict to level-1 binets, it is possible to decide in polynomial time whether there exists a binary network that displays all the binets. We also show that to find a network that displays a maximum number of the binets is NP-hard, but that there exists a simple polynomial-time 1/3-approximation algorithm for this problem. It is hoped that these results will eventually assist in the development of new methods for constructing phylogenetic networks from collections of smaller networks.

Phylogenetic networks are a generalization of phylogenetic trees that allow for representation of reticulate evolution. Recently, a space of unrooted phylogenetic networks was introduced, where such a network is a connected graph in which every vertex has degree 1 or 3 and whose leaf-set is a fixed set $X$ of taxa. This space, denoted $\mathcal{N}(X)$, is defined in terms of two operations on networks -- the nearest neighbor interchange and triangle operations -- which can be used to transform any network with leaf set $X$ into any other network with that leaf set. In particular, it gives rise to a metric $d$ on $\mathcal N(X)$ which is given by the smallest number of operations required to transform one network in $\mathcal N(X)$ into another in $\mathcal N(X)$. The metric generalizes the well-known NNI-metric on phylogenetic trees which has been intensively studied in the literature. In this paper, we derive a bound for the metric $d$ as well as a related metric $d_{N\!N\!I}$ which arises when restricting $d$ to the subset of $\mathcal{N}(X)$ consisting of all networks with $2(|X|-1+i)$ vertices, $i \ge 1$. We also introduce two new metrics on networks -- the SPR and TBR metrics -- which generalize the metrics on phylogenetic trees with the same name and give bounds for these new metrics. We expect our results to eventually have applications to the development and understanding of network search algorithms., Comment: 17 pages, 5 figures. This version has a new figure to illustrate Lemma 3.2, and a new Corollary 5.7 that bounds the distance between networks in different tiers

International audience; Binets and trinets are phylogenetic networks with two and three leaves, respectively. Here we consider the problem of deciding if there exists a binary level-1 phylogenetic network displaying a given set T of binary binets or trinets over a taxon set X , and constructing such a network whenever it exists. We show that this is NP-hard for trinets but polynomial-time solvable for binets. Moreover, we show that the problem is still polynomial-time solvable for inputs consisting of binets and trinets as long as the cycles in the trinets have size three. Finally, we present an O(3 |X | poly(|X |)) time algorithm for general sets of binets and trinets. The latter two algorithms generalise to instances containing level-1 networks with arbitrarily many leaves, and thus provide some of the first supernetwork algorithms for computing networks from a set of rooted phylogenetic networks. B Leo van Iersel

In this paper, we present a rigorous analysis on the limiting behavior of the degree distribution of the partial duplication model, a random network growth model in the duplication and divergence family that is popular in the study of biological networks. We show that for each non-negative integer k, the expected proportion of nodes of degree k approaches a limit as the network becomes large. This fills in a gap in previous studies. In addition, we prove that p=1/2, where p is the selection probability of the model, is the phase transition for the expected proportion of isolated nodes converging to 1, and hence answer a question raised in Bebek et al. [G. Bebek, P. Berenbrink, C. Cooper, T. Friedetzky, J. Nadeau, S.C. Sahinalp, The degree distribution of the generalized duplication model, Theoret. Comput. Sci. 369 (2006) 239-249]. We also obtain asymptotic bounds on the convergence rates of degree distribution. Since the observed networks typically do not contain isolated nodes, we study the subgraph consisting of all non-isolated nodes contained in the networks generated by the partial duplication model, and show that p=1/2 is again a phase transition for the limiting behavior of its degree distribution.

A realization of a finite metric space (X,d) is a weighted graph (G,w) whose vertex set contains X such that the distances between the elements of X in G correspond to those given by d. Such a realization is called optimal if it has minimal total edge weight. Optimal realizations have applications in fields such as phylogenetics, psychology, compression software and internet tomography. Given an optimal realization (G,w) of (X,d), there always exist certain ''proper'' maps from the vertex set of G into the so-called tight span of d. In [A. Dress, Trees, tight extensions of metric spaces, and the cohomological dimension of certain groups: a note on combinatorial properties of metric spaces, Adv. Math. 53 (1984) 321-402], Dress conjectured that any such map must be injective. Although this conjecture was recently disproven, in this paper we show that it is possible to characterize those optimal realizations (G,w) for which certain generalizations of proper maps-that map the geometric realization of (G,w) into the tight span instead of its vertex set-must always be injective. We also prove that these ''injective'' optimal realizations always exist, and show how they may be constructed from non-injective ones. Ultimately it is hoped that these results will contribute towards developing new ways to compute optimal realizations from tight spans.

We introduce the concept of single-linkage equivalence of edge-weighted graphs, we apply it to characterise maximal spanning trees and “ultra-similarities”, and we discuss how it relates to the popular single-linkage clustering algorithm.

In this paper we investigate guessing number, a relatively new concept linked to network coding and certain long standing open questions in circuit complexity. Here we study the bounds and a variety of properties concerning this parameter. As an application, we obtain the lower and upper bounds for shift graphs, a subclass of directed circulant graphs.

We investigate the computational complexity of finding an element of a permutation group [email protected]?S"n with minimal distance to a given @[email protected]?S"n, for different metrics on S"n. We assume that H is given by a set of generators. In particular, the size of H might be exponential in the input size, so that in general the problem cannot be solved in polynomial time by exhaustive enumeration. For the case of the Cayley Distance, this problem has been shown to be NP-hard, even if H is abelian of exponent two [R.G.E. Pinch, The distance of a permutation from a subgroup of S"n, in: G. Brightwell, I. Leader, A. Scott, A. Thomason (Eds.), Combinatorics and Probability, Cambridge University Press, 2007, pp. 473-479]. We present a much simpler proof for this result, which also works for the Hamming Distance, the l"p distance, Lee's Distance, Kendall's tau, and Ulam's Distance. Moreover, we give an NP-hardness proof for the l"~ distance using a different reduction idea. Finally, we discuss the complexity of the corresponding fixed-parameter and maximization problems.

Given a metric d on a permutation group G, the corresponding weight problem is to decide whether there exists an element g ∈ G such that d( g , e) = k for some k ∈ N . In this paper we show that this problem is NP-complete for many well known metrics.

A realization of a metric d on a finite set X is a weighted graph ( G , w ) whose vertex set contains X such that the shortest-path distance between elements of X considered as vertices in G is equal to d . Such a realization ( G , w ) is called optimal if the sum of its edge weights is minimal over all such realizations. Optimal realizations always exist, although it is NP-hard to compute them in general, and they have applications in areas such as phylogenetics, electrical networks and internet tomography. A. Dress (1984) showed that the optimal realizations of a metric d are closely related to a certain polytopal complex that can be canonically associated to d called its tight-span. Moreover, he conjectured that the (weighted) graph consisting of the zero- and one-dimensional faces of the tight-span of d must always contain an optimal realization as a homeomorphic subgraph. In this paper, we prove that this conjecture does indeed hold for a certain class of metrics, namely the class of totally-decomposable metrics whose tight-span has dimension two. As a corollary, it follows that the minimum Manhattan network problem is a special case of finding optimal realizations of two-dimensional totally-decomposable metrics.

In evolutionary biology various metrics have been defined and studied for comparing phylogenetic trees. Such metrics are used, for example, to compare competing evolutionary hypotheses or to help organize algorithms that search for optimal trees. Here we introduce a new metric dpdp on the collection of binary phylogenetic trees each labeled by the same set of species. The metric is based on the so-called parsimony score, an important concept in phylogenetics that is commonly used to construct phylogenetic trees. Our main results include a characterization of the unit neighborhood of a tree in the dpdp metric, and an explicit formula for its diameter, that is, a formula for the maximum possible value of dpdp over all possible pairs of trees labeled by the same set of species. We also show that dpdp is closely related to the well-known tree bisection and reconnection (tbr) and subtree prune and regraft (spr) distances, a connection which will hopefully provide a useful new approach to understanding properties of these and related metrics.

In evolutionary biology, biologists often face the problem of constructing a phylogenetic tree on a set $X$ of species from a multiset $\Pi$ of partitions corresponding to various attributes of these species. One approach that is used to solve this problem is to try instead to associate a tree (or even a network) to the multiset $\Sigma_{\Pi}$ consisting of all those bipartitions $\{A,X-A\}$ with $A$ a part of some partition in $\Pi$. The rational behind this approach is that a phylogenetic tree with leaf set $X$ can be uniquely represented by the set of bipartitions of $X$ induced by its edges. Motivated by these considerations, given a multiset $\Sigma$ of bipartitions corresponding to a phylogenetic tree on $X$, in this paper we introduce and study the set $P(\Sigma)$ consisting of those multisets of partitions $\Pi$ of $X$ with $\Sigma_{\Pi}=\Sigma$. More specifically, we characterize when $P(\Sigma)$ is non-empty, and also identify some partitions in $P(\Sigma)$ that are of maximum and minimum size. We also show that it is NP-complete to decide when $P(\Sigma)$ is non-empty in case $\Sigma$ is an arbitrary multiset of bipartitions of $X$. Ultimately, we hope that by gaining a better understanding of the mapping that takes an arbitrary partition system $\Pi$ to the multiset $\Sigma_{\Pi}$, we will obtain new insights into the use of median networks and, more generally, split-networks to visualize sets of partitions., Comment: 28 pages, submitted to SIAM J. Discrete Mathematics

Tree rearrangement operations typically induce a metric on the space of phylogenetic trees. One important property of these metrics is the size of the neighborhood, that is, the number of trees exactly one operation from a given tree. We present an exact expression for the size of the TBR (tree bisection and reconnection) neighborhood, thus answering a question first posed by Allen and Steel . In addition, we also obtain a characterization of the extremal trees whose TBR neighborhoods are maximized and minimized.

Given a metric d on a permutation group G, the corresponding weight problem is to decide whether there exists an element π∈G such that d(π,e)=k, for some given value k. Here we show that this problem is NP-complete for many well-known metrics. An analogous problem in matrix groups, eigenvalue-free problem, and two related problems in permutation groups, the maximum and minimum weight problems, are also investigated in this paper.