Wheeler WC and Pickett KM (2008. Topology-Bayes versus clade-Bayes in phylogenetic analysis. Mol Biol Evol. 25:447–453.) discuss two ways of summarizing the posterior probability distribution of a Bayesian phylogenetic analysis, which they refer to as "topology-Bayes" and "clade-Bayes." They claim that the clade-Bayes approach leads to problems such as "exaggerated clade support, inconsistently biased priors, and the impossibility of topology hypothesis testing," which are not problems for the topology-Bayes approach. However, their argument for topology-Bayes over clade-Bayes is based on errors in the interpretation of summary statistics associated with Bayesian phylogenetic analysis. Although there is a well-documented difference between the maximum posterior probability topology and the majority-rule consensus topology (the established terms for topology-Bayes and clade-Bayes summaries, respectively), both have a place in phylogenetic analysis. Choice of summarization strategy should be driven by choice of parameters that need to be estimated versus those to be marginalized given the evolutionary questions being asked or hypotheses being tested.

The marine cyanobacterium Prochlorococcus MED4 has the smallest sequenced genome of any photosynthetic organism. Prochlorococcus MED4 shares many genomic characteristics with chloroplasts and bacterial endosymbionts, including a reduced coding capacity, missing DNA repair genes, a minimal transcriptional regulatory network, a marked AT% bias, and an accelerated rate of amino acid changes. In chloroplasts and endosymbionts, these molecular phenotypes appear to be symptomatic of a relative increase in genetic drift due to restrictions on effective population size in the host environment. As a free-living bacterium, Prochlorococcus MED4 is not known to be subject to similar ecological constraints. To test whether the high-light-adapted Prochlorococcus MED4 is experiencing a reduction in selection efficiency resulting from genetic drift, we examine two data sets, namely, the environmental genome shotgun sequencing data from the Sargasso Sea and a set of cyanobacterial genome sequences. After integrating these data sets, we compare the evolutionary profile of a high-light Prochlorococcus group to that of a group of Synechococcus (a closely related group of marine cyanobacteria) that does not exhibit a similar small-genome syndrome. The average pairwise dN/dS ratios in the high-light-adapted Prochlorococcus group are significantly lower than those in the Synechococcus group, leading us to reject the hypothesis that the Prochlorococcus group is currently experiencing higher levels of genetic drift.

Toxoglossate marine gastropods, traditionally assigned to the families Conidae, Terebridae, and Turridae, are one of the most populous animal groups that use venom to capture their prey. These marine animals are generally characterized by a venom apparatus that consists of a muscular venom bulb and a tubular venom gland. The toxoglossan radula, often compared with a hypodermic needle for its use as a conduit to inject toxins into prey, is considered a major anatomical breakthrough that assisted in the successful initial radiation of these animals in the Cretaceous and early Tertiary. The pharmacological success of toxins from cone snails has made this group a star among biochemists and neuroscientists, but very little is known about toxins from the other Toxoglossa, and the phylogeny of these families is largely in doubt. Here we report the first molecular phylogeny for the Terebridae and use the results to infer the evolution of the venom apparatus for this group. Our findings indicate that most of the genera of terebrids are polyphyletic, and one species ("Terebra" (s.l.) jungi) is the sister group to all other terebrids. Molecular analyses combined with mapping of venom apparatus morphology indicate that the Terebridae have lost the venom apparatus at least twice during their evolution. Species in the genera Terebra and Hastula have the typical venom apparatus found in most toxoglossate gastropods, but all other terebrid species do not. For venomous organisms, the dual analysis of molecular phylogeny and toxin function is an instructive combination for unraveling the larger questions of phylogeny and speciation. The results presented here suggest a paradigm shift in the current understanding of terebrid evolution, while presenting a road map for discovering novel terebrid toxins, a largely unexplored resource for biomedical research and potential therapeutic drug development.

Several morphologically dissimilar ascomycete fungi including Schizosaccharomyces, Taphrina, Saitoella, Pneumocystis, and Neolecta have been grouped into the taxon Taphrinomycotina (Archiascomycota or Archiascomycotina), originally based on rRNA phylogeny. These analyses lack statistically significant support for the monophyly of this grouping, and although confirmed by more recent multigene analyses, this topology is contradicted by mitochondrial phylogenies. To resolve this inconsistency, we have assembled phylogenomic mitochondrial and nuclear data sets from four distantly related taphrinomycotina taxa: Schizosaccharomyces pombe, Pneumocystis carinii, Saitoella complicata, and Taphrina deformans. Our phylogenomic analyses based on nuclear data (113 proteins) conclusively support the monophyly of Taphrinomycotina, diverging as a sister group to Saccharomycotina + Pezizomycotina. However, despite the improved taxon sampling, Taphrinomycotina continue to be paraphyletic with the mitochondrial data set (13 proteins): Schizosaccharomyces species associate with budding yeasts (Saccharomycotina) and the other Taphrinomycotina group as a sister group to Saccharomycotina + Pezizomycotina. Yet, as Schizosaccharomyces and Saccharomycotina species are fast evolving, the mitochondrial phylogeny may be influenced by a long-branch attraction (LBA) artifact. After removal of fast-evolving sequence positions from the mitochondrial data set, we recover the monophyly of Taphrinomycotina. Our combined results suggest that Taphrinomycotina is a legitimate taxon, that this group of species diverges as a sister group to Saccharomycotina + Pezizomycotina, and that phylogenetic positioning of yeasts and fission yeasts with mitochondrial data is plagued by a strong LBA artifact.

Convergent evolution is a widespread phenomenon seen in diverse organisms inhabiting similar selective environments. However, it is unclear if similar phenotypes are produced by the same or different genes and mutations. Here we analyze the molecular mechanisms underlying convergent pigment pattern among subspecies of the beach mouse (Peromyscus polionotus) inhabiting the Gulf and Atlantic coasts of Florida. In these two geographic regions, separated by more than 300 km, "beach mice" have lighter colored coats than do their mainland counterparts, produced by natural selection for camouflage against the pale coastal sand dunes. We measured color pattern in eight beach mouse subspecies and showed that three of the Gulf Coast subspecies are more phenotypically similar to an Atlantic coast subspecies than to their Gulf Coast neighbors. However, light-colored beach mice do not form a monophyletic group. Previous results implicated a single derived amino acid change in the melanocortin-1 receptor (Mc1r) as a major contributor to pigment pattern in the Gulf Coast beach mice; despite phenotypic similarities, the derived Mc1r allele was not found in the Atlantic coast beach mouse populations. Here we show that Atlantic coast beach mice have high levels of Mc1r polymorphism but they lack unique alleles. Functional assays revealed that single amino acid mutations segregating in Atlantic coast beach mice do not cause any change in Mc1r activity compared with the activity of Mc1r from dark-colored mice. These joint results show that convergent pigment patterns in recently diverged beach mouse subspecies—whose developmental constraints are presumably similar—have evolved through a diversity of genetic mechanisms.

Two rounds of whole-genome duplications are thought to have played an important role in the establishment of gene repertoires in vertebrates. These events occurred during chordate evolution after the split of the urochordate and cephalochordate lineages but before the radiation of extant gnathostomes (jawed vertebrates). During this interval, diverse agnathans (jawless fishes), including cyclostomes (hagfishes and lampreys), diverged. However, there is no solid evidence for the timing of these genome duplications in relation to the divergence of cyclostomes from the gnathostome lineage. We conducted cDNA sequencing in diverse early vertebrates for members of homeobox-containing (Dlx and ParaHox) and other gene families that would serve as landmarks for genome duplications. Including these new sequences, we performed a molecular phylogenetic census using the maximum likelihood method for 55 gene families. In most of these gene families, we detected many more gene duplications before the cyclostome–gnathostome split, than after. Many of these gene families (e.g., visual opsins, RAR, Notch) have multiple paralogs in conserved, syntenic genomic regions that must have been generated by large-scale duplication events. Taken together, this indicates that the genome duplications occurred before the cyclostome–gnathostome split. We propose that the redundancy in gene repertoires possessed by all vertebrates, including hagfishes and lampreys, was introduced primarily by genome duplications. Apart from subsequent lineage-specific modifications, these ancient genome duplication events might serve generally to distinguish vertebrates from invertebrates at the genomic level.

In Drosophila, odorant receptors are encoded by an old and moderately sized multigene family. Or22a and Or22b are two tandemly arranged genes of this family that have proved to be the result of a rather young duplication. Nucleotide variation in the region spanning both duplicates was surveyed in four natural populations (two African and two non-African) of Drosophila melanogaster and also analyzed in species of the melanogaster subgroup. The intraspecific survey revealed a particular copy-number polymorphism in some of the studied populations, with the two genes (Or22a and Or22b) present in the long variant and a single chimeric gene (Or22ab) present in the short variant. Estimated nucleotide diversity was higher in the short than in the long variant, despite the ancestral character of the latter variant in D. melanogaster. The general skew toward low-frequency variants detected in the non-African long variant and its reduced level of silent polymorphism relative to divergence is consistent with the recent fixation of an advantageous mutation at, or nearby, the Or22 long variant region. The nonnegligible frequency of the short variant and the presence of a highly divergent haplotype in the East African sample would point to direct or indirect selection for its maintenance in the species. There was evidence for a generally more rapid evolution of the Or22b copy at both synonymous and nonsynonymous sites. However, an excess of nonsynonymous substitutions was only detected in the early history of this copy.

The Adhesion G-protein–coupled receptors (GPCRs) are the most complex gene family among GPCRs with large genomic size, multiple introns, and a fascinating flora of functional domains, though the evolutionary origin of this family has been obscure. Here we studied the evolution of all class B (7tm2)–related genes, including the Adhesion, Secretin, and Methuselah families of GPCRs with a focus on nine genomes. We found that the cnidarian genome of Nematostella vectensis has a remarkably rich set of Adhesion GPCRs with a broad repertoire of N-terminal domains although this genome did not have any Secretin GPCRs. Moreover, the single-celled and colony-forming eukaryotes Monosiga brevicollis and Dictyostelium discoideum contain Adhesion-like GPCRs although these genomes do not have any Secretin GPCRs suggesting that the Adhesion types of GPCRs are the most ancient among class B GPCRs. Phylogenetic analysis found Adhesion group V (that contains GPR133 and GPR144) to be the closest relative to the Secretin family in the Adhesion family. Moreover, Adhesion group V sequences in N. vectensis share the same splice site setup as the Secretin GPCRs. Additionally, one of the most conserved motifs in the entire Secretin family is only found in group V of the Adhesion family. We suggest therefore that the Secretin family of GPCRs could have descended from group V Adhesion GPCRs. We found a set of unique Adhesion-like GPCRs in N. vectensis that have long N-termini containing one Somatomedin B domain each, which is a domain configuration similar to that of a set of Adhesion-like GPCRs found in Branchiostoma floridae. These sequences show slight similarities to Methuselah sequences found in insects. The extended class B GPCRs have a very complex evolutionary history with several species-specific expansions, and we identified at least 31 unique N-terminal domains originating from other protein classes. The overall N-terminal domain structure, however, concurs with the phylogenetic analysis of the transmembrane domains, thus enabling us to track the origin of most of the subgroups.

Crucifers (Brassicaceae, Cruciferae) are a large family comprising some 338 genera and c. 3,700 species. The family includes important crops as well as several model species in various fields of plant research. This paper reports new genome size (GS) data for more than 100 cruciferous species in addition to previously published C-values (the DNA amount in the unreplicated gametic nuclei) to give a data set comprising 185 Brassicaceae taxa, including all but 1 of the 25 tribes currently recognized. Evolution of GS was analyzed within a phylogenetic framework based on gene trees built from five data sets (matK, chs, adh, trnLF, and ITS). Despite the 16.2-fold variation across the family, most Brassicaceae species are characterized by very small genomes with a mean 1C-value of 0.63 pg. The ancestral genome size (ancGS) for Brassicaceae was reconstructed as anc1C = 0.50 pg. Approximately 50% of crucifer taxa analyzed showed a decrease in GS compared with the ancGS. The remaining species showed an increase in GS although this was generally moderate, with significant increases in C-value found only in the tribes Anchonieae and Physarieae. Using statistical approaches to analyze GS, evolutionary gains or losses in GS were seen to have accumulated disproportionately faster within longer branches. However, we also found that GS has not changed substantially through time and most likely evolves passively (i.e., a tempo that cannot be distinguished between neutral evolution and weak forms of selection). The data reveal an apparent paradox between the narrow range of small GSs over long evolutionary time periods despite evidence of dynamic genomic processes that have the potential to lead to genome obesity (e.g., transposable element amplification and polyploidy). To resolve this, it is suggested that mechanisms to suppress amplification and to eliminate amplified DNA must be active in Brassicaceae although their control and mode of operation are still poorly understood.

The mitochondrial genome of grape (Vitis vinifera), the largest organelle genome sequenced so far, is presented. The genome is 773,279 nt long and has the highest coding capacity among known angiosperm mitochondrial DNAs (mtDNAs). The proportion of promiscuous DNA of plastid origin in the genome is also the largest ever reported for an angiosperm mtDNA, both in absolute and relative terms. In all, 42.4% of chloroplast genome of Vitis has been incorporated into its mitochondrial genome. In order to test if horizontal gene transfer (HGT) has also contributed to the gene content of the grape mtDNA, we built phylogenetic trees with the coding sequences of mitochondrial genes of grape and their homologs from plant mitochondrial genomes. Many incongruent gene tree topologies were obtained. However, the extent of incongruence between these gene trees is not significantly greater than that observed among optimal trees for chloroplast genes, the common ancestry of which has never been in doubt. In both cases, we attribute this incongruence to artifacts of tree reconstruction, insufficient numbers of characters, and gene paralogy. This finding leads us to question the recent phylogenetic interpretation of Bergthorsson et al. (2003, 2004) and Richardson and Palmer (2007) that rampant HGT into the mtDNA of Amborella best explains phylogenetic incongruence between mitochondrial gene trees for angiosperms. The only evidence for HGT into the Vitis mtDNA found involves fragments of two coding sequences stemming from two closteroviruses that cause the leaf roll disease of this plant. We also report that analysis of sequences shared by both chloroplast and mitochondrial genomes provides evidence for a previously unknown gene transfer route from the mitochondrion to the chloroplast.

Simple sequence repeats (SSRs) are very common short repeats in eukaryotic genomes. "Long" SSRs are considered "hypermutable" sequences because they exhibit a high rate of expansion and contraction. Because they are potentially deleterious, long SSRs tend to be uncommon in coding sequences. However, several genes contain long SSRs in their exonic sequences. Here, we identify 1,291 human genes that host a mononucleotide SSR long enough to be prone to expansion or contraction, being called hypermutable hereafter. On the basis of Gene Ontology annotations, we show that only a restricted number of functions are overrepresented among those hypermutable genes including cell cycle and maintenance of DNA integrity. Using a probabilistic model, we show that genes involved in these functions are expected to host long SSRs because they tend to be long and/or are biased in nucleotide composition. Finally, we show that for almost all functions we observe fewer hypermutable sequences than expected under a neutral model. There are however interesting exceptions, for example, genes involved in protein and RNA transport, as well as meiosis and mismatch repair functions that have as many hypermutable genes as expected under neutrality. Conversely, there are functions (e.g., collagen-related genes) where hypermutable genes are more often avoided than in other functions. Our results show that, even though several functions harbor unusually long SSR in their exons, long SSRs are deleterious sequences in almost all functions and are removed by purifying selection. The strength of this purifying selection however greatly varies from function to function. We discuss possible explanations for this intriguing result.

There are now many known cases of orthologous or unrelated proteins in different species that have undergone parallel evolution to satisfy a similar function. However, there are no reported cases of parallel evolution for proteins that bind a common ligand but have different functions. We focused on two proteins that have different functions in steroid hormone biosynthesis and action but bind a common ligand, androgen. The first protein, androgen receptor (AR), is a nuclear hormone receptor and the second one, aromatase (cytochrome P450 19 [CYP19]), converts androgen to estrogen. We hypothesized that binding of the androgen ligand has exerted common selective pressure on both AR and CYP19, resulting in a signature of parallel evolution between these two proteins, though they perform different functions. Consistent with this hypothesis, we found that rates of amino acid change in AR and CYP19 are strongly correlated across the metazoan phylogeny, whereas no significant correlation was found in the control set of proteins. Moreover, we inferred that genomic toolkits required for steroid biosynthesis and action were present in a basal metazoan, cnidarians. The close similarities between vertebrate and sea anemone AR and CYP19 suggest a very ancient origin of their endocrine functions at the base of metazoan evolution. Finally, we found evidence supporting the hypothesis that the androgen-to-estrogen ratio determines the gonadal sex in all metazoans.

If substitution rates are not the same on the two complementary DNA strands, a substitution is considered strand asymmetric. Such substitutional strand asymmetries are determined here for the three most frequent types of substitution on the human genome (C -> T, A -> G, and G -> T). Substitution rate differences between both strands are estimated for 4,590 human genes by aligning all repeats occurring within the introns with their ancestral consensus sequences. For 1,630 of these genes, both coding strand and noncoding strand rates could be compared with rates in gene-flanking regions. All three rates considered are found to be on average higher on the coding strand and lower on the transcribed strand in comparison to their values in the gene-flanking regions. This finding points to the simultaneous action of rate-increasing effects on the coding strand—such as increased adenine and cytosine deamination—and transcription-coupled repair as a rate-reducing effect on the transcribed strand. The common behavior of the three rates leads to strong correlations of the rate asymmetries: Whenever one rate is strand biased, the other two rates are likely to show the same bias. Furthermore, we determine all three rate asymmetries as a function of time: the A -> G and G -> T rate asymmetries are both found to be constant in time, whereas the C -> T rate asymmetry shows a pronounced time dependence, an observation that explains the difference between our results and those of an earlier work by Green et al. (2003. Transcription-associated mutational asymmetry in mammalian evolution. Nat Genet. 33:514–517.). Finally, we show that in addition to transcription also the replication process biases the substitution rates in genes.

Hantaviruses are considered one of the best examples of a long-term association between RNA viruses and their hosts. Based on the appearance of strong host specificity, it has been suggested that hantaviruses cospeciated with the rodents and insectivores they infect since these mammals last shared a common ancestor, approximately 100 million years ago. We tested this hypothesis of host–virus codivergence in two ways: 1) we used cophylogenetic reconciliation analysis to assess the fit of the virus tree onto that of the host and 2) we estimated the evolutionary rates and divergence times for the Hantavirus genus using a Bayesian Markov Chain Monte Carlo method and similarly compared these with those of their hosts. Our reconciliation analysis provided no evidence for a history of codivergence between hantaviruses and their hosts. Further, the divergence times for the Hantavirus genus were many orders of magnitude too recent to correspond with the timescale of their hosts' speciation. We therefore propose that apparent similarities between the phylogenies of hantaviruses and their mammalian hosts are the result of a more recent history of preferential host switching and local adaptation. Based on the presence of clade-defining amino acids in all genomic segments, we propose that the patterns of amino acid replacement in these viruses are also compatible with a history of host-specific adaptation.

In the past 40 years, there has been increasing acceptance that variation in levels of gene expression represents a major source of evolutionary novelty. Gene expression divergence is therefore likely to be involved in the emergence of incipient species, namely, in a context of adaptive radiation. In the lake whitefish species complex (Coregonus clupeaformis), previous microarray experiments have led to the identification of candidate genes potentially implicated in the parallel evolution of the limnetic dwarf lake whitefish, which is highly distinct from the benthic normal lake whitefish in life history, morphology, metabolism, and behavior, and yet diverged from it only ~15,000 years before present. The aim of the present study was to address transcriptional divergence for six candidate genes among lake whitefish and European whitefish (Coregonus lavaretus) species pairs, as well as lake cisco (Coregonus artedi) and vendace (Coregonus albula). The main goal was to test the hypothesis that parallel phenotypic adaptation toward the use of the limnetic niche in coregonine fishes is accompanied by parallelism in candidate gene transcription as measured by quantitative real-time polymerase chain reaction. Results obtained for three candidate genes, whereby parallelism in expression was observed across all whitefish species pairs, provide strong support for the hypothesis that divergent natural selection plays an important role in the adaptive radiation of whitefish species. However, this parallelism in expression did not extend to cisco and vendace, thereby infirming transcriptional convergence between limnetic whitefish species and their limnetic congeners for these genes. As recently proposed (Lynch 2007a. The evolution of genetic networks by non-adaptive processes. Nat Rev Genet. 8:803–813), these results may suggest that convergent phenotypic evolution can result from nonadaptive shaping of genome architecture in independently evolved coregonine lineages.

According to current taxonomical rules, a bona fide bacterial species is a genomic species characterized by the genomic similarity of its members. It has been proposed that the genomic cohesion of such clusters may be related to sexual isolation, which limits gene flow between too divergent bacteria. Homologous recombination is one of the most studied mechanisms responsible for this genetic isolation. Previous studies on several bacterial models showed that recombination frequencies decreased exponentially with increasing DNA sequence divergence. In the present study, we investigated this relationship in the Agrobacterium tumefaciens species complex, which allowed us to focus on sequence divergence in the vicinity of the genetic boundaries of genomic species. We observed that the sensitivity of the recombination frequency to DNA divergence fitted a log-linear function until approximately 10% sequence divergence. The results clearly revealed that there was no sharp drop in recombination frequencies at the point where the sequence divergence distribution showed a "gap" delineating genomic species. The ratio of the recombination frequency in homogamic conditions relative to this frequency in heterogamic conditions, that is, sexual isolation, was found to decrease from 8 between the most distant strains within a species to 9 between the most closely related species, for respective increases from 4.3% to 6.4% mismatches in the marker gene chvA. This means that there was only a 1.13-fold decrease in recombination frequencies for recombination events at both edges of the species border. Hence, from the findings of this investigation, we conclude that—at least in this taxon—sexual isolation based on homologous recombination is likely not high enough to strongly hamper gene flow between species as compared with gene flow between distantly related members of the same species. The 70% relative binding ratio cutoff used to define bacterial species is likely correlated to only minor declines in homologous recombination frequencies. Consequently, the sequence diversity, as a mechanistic factor for the efficiency of recombination (as assayed in the laboratory), appears to play little role in the genetic cohesion of bacterial species, and thus, the genomic species definition for prokaryotes is definitively not reconcilable with the biological species concept for eukaryotes.

Avian influenza A viruses (AIVs), including the H5N1, H9N2, and H7N7 subtypes, have been directly transmitted to humans, raising concerns over the possibility of a new influenza pandemic. To prevent a future avian influenza pandemic, it is very important to fully understand the molecular basis driving the change in AIV virulence and host tropism. Although virulent variants of other viruses have been generated by homologous recombination, the occurrence of homologous recombination within AIV segments is controversial and far from proven. This study reports three circulating H9N2 AIVs with similar mosaic PA genes descended from H9N2 and H5N1. Additionally, many homologous recombinants are also found deposited in GenBank. Recombination events can occur in PB2, PB1, PA, HA, and NP segments and between lineages of the same/different serotype. These results collectively demonstrate that intragenic recombination plays a role in driving the evolution of AIVs, potentially resulting in effects on AIV virulence and host tropism changes.

During insemination, males of internally fertilizing species transfer a complex array of seminal fluid proteins to the female reproductive tract. These proteins can have profound effects on female reproductive physiology and behavior and are thought to mediate postcopulatory sexual selection and intersexual conflict. Such selection may cause seminal fluid to evolve rapidly, with potentially important consequences for speciation. Here we investigate the evolution of seminal fluid proteins in a major mammalian radiation, the muroid rodents, by quantifying diversity in seminal fluid proteome composition for the first time across a broad range of closely related species. Using comparative proteomics techniques to identify and cross-match proteins, we demonstrate that rodent seminal fluid is highly diverse at the level of both proteomes and individual proteins. The striking interspecific heterogeneity in seminal fluid composition revealed by our survey far exceeds that seen in a second proteome of comparable complexity, skeletal muscle, indicating that the complement of proteins expressed in seminal fluid may be subject to rapid diversification. We further show that orthologous seminal fluid proteins exhibit substantial interspecific variation in molecular mass. Because this variation cannot be attributed to differential glycosylation or radical differences in termination sites, it is strongly suggestive of rapid amino acid divergence. Sperm competition is implicated in generating such divergence for at least one major seminal fluid protein in our study, SVS II, which is responsible for copulatory plug formation via transglutaminase-catalyzed cross-linking after insemination. We show that the molecular mass of SVS II is positively correlated with relative testis size across species, which could be explained by selection for an increased number of cross-linking sites involved in the formation of the copulatory plug under sperm competition.

Despite their ability to interbreed and produce fertile offspring, there is continued disagreement about the genetic relationship of the domestic horse (Equus caballus) to its endangered wild relative, Przewalski's horse (Equus przewalskii). Analyses have differed as to whether or not Przewalski's horse is placed phylogenetically as a separate sister group to domestic horses. Because Przewalski's horse and domestic horse are so closely related, genetic data can also be used to infer domestication-specific differences between the two. To investigate the genetic relationship of Przewalski's horse to the domestic horse and to address whether evolution of the domestic horse is driven by males or females, five homologous introns (a total of ~3 kb) were sequenced on the X and Y chromosomes in two Przewalski's horses and three breeds of domestic horses: Arabian horse, Mongolian domestic horse, and Dartmoor pony. Five autosomal introns (a total of ~6 kb) were sequenced for these horses as well. The sequences of sex chromosomal and autosomal introns were used to determine nucleotide diversity and the forces driving evolution in these species. As a result, X chromosomal and autosomal data do not place Przewalski's horses in a separate clade within phylogenetic trees for horses, suggesting a close relationship between domestic and Przewalski's horses. It was also found that there was a lack of nucleotide diversity on the Y chromosome and higher nucleotide diversity than expected on the X chromosome in domestic horses as compared with the Y chromosome and autosomes. This supports the hypothesis that very few male horses along with numerous female horses founded the various domestic horse breeds. Patterns of nucleotide diversity among different types of chromosomes were distinct for Przewalski's in contrast to domestic horses, supporting unique evolutionary histories of the two species.

Noncoding RNAs (ncRNAs) are transcripts that do not code for protein but rather function as RNA in catalytic, regulatory, or structural roles in the cell. ncRNAs are involved in universally conserved biological processes, including protein synthesis and gene regulation, and have more specific roles, such as in X-chromosome inactivation in eutherian mammals. In this paper, we propose and investigate a hypothesis for patterns of sequence selection in structurally conserved ncRNAs. Previous attempts at defining RNA selection compared rates of evolution between paired and unpaired bases with largely inconclusive results. Our approach focuses only on paired bases in ncRNAs with conserved structure. By analogy to the different properties of codon positions based on the genetic code, we use a well-developed energy model for RNA structure to classify stem positions into structural classes and argue that they are under different selective constraints. We validate the hypothesis on several RNA families and use simulated data to verify the evolutionary origin of signals. Our class labeling is shown to be a better model of ncRNA evolution than the tradition of treating stem positions equally. As well as providing a better understanding of RNA evolution, the evolutionary footprint we identify can easily be incorporated into gene finders to improve their specificity.