Present biodiversity has come about through processes of diversification and extinction of species, and the mechanisms that drive these processes are a central focus in evolutionary biology (e.g. [1–4]). One intriguing relationship that has been revealed through studies of branch lengths on molecular phylogenies is a link between the rate of molecular evolution and the net diversification rate. A correlation between evolutionary rates and species diversity has been found in several groups including flowering plants [5, 6], reptiles , birds [7, 8], and other metazoan phyla, orders, and classes .
However, not all the datasets analysed have provided evidence for a link between diversification rates and rates of molecular evolution. One study of a large number of phylogenies found a relationship between root-to-tip branch lengths and net diversification in around half of the phylogenies tested, but it is not clear whether this was due to low power or lack of a common trend . Another study examined genetic data within the mammals and found no evidence of an association between molecular rates and net diversification . Accordingly, the universality and causes of the link remain uncertain.
There are three possible causes for the association between net diversification and the rate of molecular evolution [5, 6]. One is that the process of diversification drives changes in the rate of molecular evolution. Speciation might influence the rate of molecular evolution through positive selection on particular genes associated with adaptation to novel niches . Speciation could also cause genome-wide increases in substitution rate if speciation is typically associated with population subdivision [13, 14]. This is because a reduction in effective population size (N
) can cause a higher rate of fixation of nearly neutral mutations (e.g. ), leading to a faster substitution rate .
Conversely, a higher rate of molecular evolution may increase the diversification rate. A faster mutation rate may hasten differentiation between populations and promote reproductive incompatibility . For example, it has been suggested that higher standing genetic diversity in populations at low latitude may contribute to faster diversification in the tropics . Increased standing genetic variation may produce more raw material for adaptation  or reduce the likelihood of extinction . However, a recent study of orchids found no evidence for a link between population genetic variability and net diversification rate . A higher rate of molecular evolution may increase the rate of diversification by accelerating the formation of hybrid incompatibility, occurring through the accumulation of genetic incompatibilities between the genomes of the diverging populations .
Alternatively, there might be a third factor that influences both the rate of molecular evolution and diversification rate, creating an indirect link between diversification and molecular evolution. For example, environmental energy (temperature and UV light) has been associated with both the diversification rate and the rate of molecular evolution in angiosperms . Other potential third factors are life history features, such as size or generation length, which are linked with the rate of molecular evolution and diversification rates of angiosperms and several metazoan taxa [9, 22–24]. It has also been suggested that both morphological and molecular rates of change may be connected to diversification rate . Whether the correlation between rate of molecular evolution and net diversification has a causal or indirect effect needs more investigation.
One way of disentangling the potential causes of the observed relationship between diversification rate and rate of molecular evolution is to partition substitutions in protein-coding genes into synonymous and non-synonymous substitutions. Synonymous mutations do not change the amino acid sequence of a protein and hence are expected to behave as neutral. If so, then the synonymous substitution rate (dS) should reflect only the mutation rate . Nonsynonymous mutations are expected to have a range of fitness effects, including neutral, positive and negative, so may be subject to both drift and selection. An increase in the nonsynonymous substitution rate (dN) relative to the synonymous rate (dS) can occur through positive selection promoting the fixation of nonsynonymous mutations, or through a reduction in population size increasing the rate of fixation of nearly neutral mutations by drift.
The link between rate of molecular evolution and diversification rate has been attributed to the action of selection during speciation, or to a reduction in average population size in lineages undergoing diversification , both of which would be expected to increase the relative rate of nonsynonymous substitutions. However, studies in angiosperms , reptiles , and birds [7, 8] have found a correlation between synonymous substitutions and net diversification, leading to the suggestion that the link between molecular rates and net diversification may be driven by the mutation rate.
Here, we focus on the rate of molecular evolution in chloroplast genes. Genetic changes in chloroplast genomes have been implicated in the process of speciation in plants. Coevolution between organelle and nuclear genomes has been recognized as an important factor in plant diversification . Plastome-genome incompatibility can cause hybrid sterility or inviability, by disrupting sexual reproduction, leaf morphologies, and machineries for photosynthesis or respiration [27–29]. Some of the genetic events in chloroplasts that produce these aberrations are gene duplications, loss of gene complexes and genome rearrangements [26, 30, 31]. The resulting incompatibilities are probably generalized phenomena in plants, and the evolutionary consequence is that they can enhance post-zygotic barriers during speciation [26, 29, 32, 33]. It seems possible, then, that variation in rates of molecular evolution of chloroplasts could also influence the speed of genetic isolation, and hence the diversification rate of plant lineages.
Using a phylogenetic comparative analysis of sister pairs , we investigated the relationship between rates of molecular evolution and net diversification in chloroplast genes of the plant family Proteaceae. This highly diverse family is mostly restricted to the Southern Hemisphere. It contains 79 recognized genera and around 1600 species, and some of its most diverse groups are the Australian genus Banksia and the African genus Protea. The high diversity of Proteaceae makes it a particularly attractive case study for diversification (e.g. [35–37]). In addition, the family has stark contrasts in species-richness between genera even within its biodiversity hotspots . Of particular interest to this study are the numerous cases of monophyletic sister clades with remarkable differences in number of species. For example, the genus Protea has 112 species, while its sister genus Faurea has 15, and the Banksia lineage (including the dryandras) has 169 species while its sister lineage of the genera Austromuellera and Musgravea contains only 4.
We focus on the rates of evolution of six chloroplast genes available for a genus level phylogeny of the family Proteaceae . We use three protein-coding genes to estimate and contrast rates of synonymous (dS) and non-synonymous (dN) substitutions. Comparing dN, dS, and ω (dN/dS) to species-richness of clades allows us to separate the effect of mutation rate on net diversification from the effect of selection and effective population size. In this way, we aim to provide insight into the factors underlying the correlation between rates of molecular evolution and net diversification.