Multiple chromosomal rearrangements structured the ancestral vertebrate Hox-bearing protochromosomes.

Vincent J Lynch,Günter P Wagner,Takashi Gojobori

doi:10.1371/journal.pgen.1000349

Vincent J Lynch, Günter P Wagner + Show 1 more

Open Access

https://doi.org/10.1371/journal.pgen.1000349

Copy DOI

Journal: PLoS Genetics	Publication Date: Jan 23, 2009
Citations: 74	License type: CC BY 4.0

Affiliation: Yale University

Abstract

While the proposal that large-scale genome expansions occurred early in vertebrate evolution is widely accepted, the exact mechanisms of the expansion—such as a single or multiple rounds of whole genome duplication, bloc chromosome duplications, large-scale individual gene duplications, or some combination of these—is unclear. Gene families with a single invertebrate member but four vertebrate members, such as the Hox clusters, provided early support for Ohno's hypothesis that two rounds of genome duplication (the 2R-model) occurred in the stem lineage of extant vertebrates. However, despite extensive study, the duplication history of the Hox clusters has remained unclear, calling into question its usefulness in resolving the role of large-scale gene or genome duplications in early vertebrates. Here, we present a phylogenetic analysis of the vertebrate Hox clusters and several linked genes (the Hox “paralogon”) and show that different phylogenies are obtained for Dlx and Col genes than for Hox and ErbB genes. We show that these results are robust to errors in phylogenetic inference and suggest that these competing phylogenies can be resolved if two chromosomal crossover events occurred in the ancestral vertebrate. These results resolve conflicting data on the order of Hox gene duplications and the role of genome duplication in vertebrate evolution and suggest that a period of genome reorganization occurred after genome duplications in early vertebrates.

Highlights

Ohno’s hypothesis [1] that two-rounds of whole genome duplication occurred in the ancestor of extant vertebrates, over 450 million years ago (Figure 1A), has generally gained wide acceptance
Bayesian inference (BI) and maximum likelihood (ML) methods overcome these limitations by being based on an explicit model of nucleotide substitution that accounts for variation in evolutionary rates between nucleotide sites, but differ in how branch supports are assessed [27]
BI uses the posterior distribution of trees sampled during the tree search to indicate branch support and reflect the probability the branch is correctly inferred given the data and the model; posterior probabilities generated from BI more accurately reflect branch support, but can be prone to over estimate confidence in clade support [27]

Summary

Introduction

Ohno’s hypothesis [1] that two-rounds of whole genome duplication (the 2R-model) occurred in the ancestor of extant vertebrates, over 450 million years ago (Figure 1A), has generally gained wide acceptance. Even though there is wide support for 2R-model, the evidence for it is still conflicting. Central to this debate has been the duplication history of Hox clusters and associated linked genes (the ‘‘Hox paralogon’’, Figure 1A/B). There is only a single cluster with associated linked genes in invertebrates [20]. This 1:4 ratio has been used to support the most widely held version of the 2R-model in which two rounds of whole genome duplications were followed by extensive gene loss

Methods

Results

Discussion

Conclusion