The Incompatible Desiderata of Gene Cluster Properties

Abstract

Abstract. There is widespread interest in comparative genomics in determining if historically and/or functionally related genes are spatially clustered in the genome, and whether the same sets of genes reappear in clusters in two or more genomes. We formalize and analyze the desirable properties of gene clusters and cluster definitions. Through detailed analysis of two commonly applied types of cluster, r-windows and maxgap, we investigate the extent to which a single definition can embody all of these properties simultaneously. We show that many of the most important properties are difficult to satisfy within the same definition. We also examine whether one commonly assumed property, which we call nestedness, is satisfied by the structures present in real genomic data.

Analysis of Gene Order Evolution Beyond Single-Copy Genes

“…From a combinatorial point of view, various formal models of conserved blocks of genes, also called gene clusters or synteny blocks of genes have been introduced [8,40,65]. In particular, the notion of common intervals is a first generalization of conserved segments in which we relax the conditions that genes appear in the same order or the same orientation.…”

Issues in the Reconstruction of Gene Order Evolution

“…There are a number of formal criteria for gene clustering in two or more organisms, giving rise to cluster detection algorithms and statistical tests for the significance of clusters. These methods, comprehensively reviewed by Hoberman and Durand [65] , all depend on arbitrary parameters that control, in different ways, the number of genes and the proximity of these genes on the chromosome in order to be considered a cluster.…”

Diagnosing duplications – can it be done?

“…However, there are no accepted, formal specifications for such regions, nor is there an explicit discussion of the properties that are desirable in such a specification. As a first step, we have proposed properties whereby paralogon definitions can be compared and evaluated [63]: (i) Order: a paralogon is completely ordered if the order of the paralogs in one region is either identical to or the exact inverse of the order in the other region. (ii) Nestedness: a paralogon of size k is nested if it contains within it a paralogon of every size !k (i.e.…”