The classic view of evolution as "shifting gene frequencies" in the Modern Synthesis literally means that evolution is the modulation of existing variation ("standing variation"), as opposed to a "new mutations" view of evolution as a 2-step process of mutational origin followed by acceptance-or-rejection (via selection and drift). The latter view has received renewed attention, yet its implications for evolutionary causation still are not widely understood. We review theoretical results showing that this conception of evolution allows for a role of mutation as a cause of nonrandomness, a role that could be important but has been misconceived and associated misleadingly with neutral evolution. Specifically, biases in the introduction of variation, including mutational biases, may impose predictable biases on evolution, with no necessary dependence on neutrality. As an example of how important such effects may be, we present a new analysis partitioning the variance in mean rates of amino acid replacement during human-chimpanzee divergence to components of codon mutation and amino acid exchangeability. The results indicate that mutational effects are not merely important but account for most of the variance explained. The challenge that such results pose for comparative genomics is to address mutational effects as a necessary part of any analysis of causal factors. To meet this challenge requires developing knowledge of mutation as a biological process, understanding how mutation imposes propensities on evolution, and applying methods of analysis that incorporate mutational effects.

Synonymous codon usage bias has been extensively studied in Bacteria and Eukaryotes and yet there has been little investigation in the third domain of life, the Archaea. In this thesis I therefore examine the coding sequences of nearly 70 species of Archaea to explore patterns of codon bias. Heterogeneity in codon usage among genes was initially explored for a single species, Methanococcus maripaludis, where patterns were explained by a single major trend associated with expression level and attributed to natural selection. Unlike the bacterium Escherichia coli, selection was largely restricted to two-fold degenerate sites. Analyses of patterns of codon usage bias within genomes were extended to the other species of Archaea, where variation was more commonly explained by heterogeneity in G+C content and asymmetric base composition. By comparison with bacterial genomes, far fewer trends were found to be associated with expression level, implying a reduced prevalence of translational selection among Archaea. The strength of selected codon usage bias (S) was estimated for 67 species of Archaea, and revealed that natural selection has had less impact in shaping patterns of codon usage across Archaea than across many species of Bacteria. Variation in S was explained by the combined effects of growth rate and optimal growth temperature, with species growing at high temperatures exhibiting weaker than expected selection given growth rate. Such a relationship is expected if temperature kinetically modulates growth rate via its impact upon translation elongation, since rapid elongation rates at high temperatures reduce the selective benefit of optimal codon usage for the efficiency of translation. Consistent with this, growth temperature is negatively correlated with minimal generation time, and numbers of rRNA operons and tRNA genes are reduced at high growth temperatures. The large fraction of thermophilic Archaea relative to Bacteria account for the lower values of S observed. Two major trends were found to describe variation in codon usage among archaeal genomes; the first was attributed to GC3s and the second was associated with arginine codon usage and was linked both with growth temperature and the genome-wide excess of G over C content. The latter is unlikely to reflect thermophilic adaptation since the codon primarily underlying the trend appears to be selectively disfavoured. No correlations were observed with genome wide GC3s and optimal growth temperature and neither was GC3s associated with aerobiosis. The identities of optimal codons were explored and found to be invariant across U and C-ending two-fold degenerate amino acid groups. The identity of optimal codons and anticodons across four and six-fold degenerate amino acid groups was found to vary with mutational bias. As was first observed in M. maripaludis, selected codon usage bias was consistently greater across two-fold relative to four-fold degenerate amino acid groups across Archaea. This broad pattern could reflect ancestral patterns of optimal codon divergence, prevalent among four-fold but not two-fold degenerate amino acid groups. Consistent with this, the strength of selected codon usage bias was found to be reduced following the divergence of optimal codons, and implies that optimal codon divergence typically proceeds following the relaxation of selection. Finally, a method was developed to partition the strength of selection (S) into separate components reflecting selection for translational efficiency (Seff) and selection for translational accuracy (Sacc) by comparing the codon usage across conserved and nonconserved amino acid residues. While estimates of Sacc are somewhat sensitive to the designation of conserved sites, a general pattern emerged whereby accuracy-selected codon usage bias was consistently strongest across a subset of the most highly conserved sites. Several estimates of Sacc were consistently higher than the 95% range of null values regardless of the dataset, providing evidence for accuracy-selected codon usage bias in these species.

By comparing sequence diversity and divergence on sex chromosomes one can study how the rate of evolution in affected by mutation and/or selection. The rate of mutation in male biased, meaning that relatively more mutations are created in the male germ line than in the female. Since the male mutation bias (αm) most likely is a consequence of the difference in cell divisions between male and female germ lines, life history characters that affect this difference should covary with αm. Indeed, we found a positive correlation between estimates of αm and increased generation times and increased intensity of sperm competition. We have also found that estimates of αm varied significantly between gametologous introns located on the sex chromosomes. This could be a consequence of the variation in substitution rates between loci. Population genetics theory predicts that both positive and negative selection reduce genetic diversity around a selected locus at a distance determined by the rate of recombination. Consequently, a non-recombining chromosome, like the female specific W chromosome in birds, selection is expected to have a large impact on sequence diversity. Indeed, in a large sequence screening we found only one segregating site among 7643 base pairs sequenced in 47 chicken females. Furthermore, we also found that deleterious substitutions are fixed in a higher rate for W- than Z-linked sequences, which is in agreement with the lack of recombination and strong genetic drift due to the low effective population size. Rarely non-synonymous mutations are beneficial for an individual, but when it happens, the mutation is positively selected and rapidly reaches fixation in a population. We have found that positive selection has been acting on the female reproductive protein, zona pellucida c in birds. This rapid evolution is likely a mechanism to prevent hybridisation.

Bacteriophages (hereafter referred as phages) can translate their mRNAs efficiently by maximizing the use of codons decoded by the most abundant tRNAs of their bacterial hosts. Translation efficiency directly influences phage fitness and evolution. Reengineered phages find application in controlling their host population in both health and industry. The objective of this thesis work is to examine the factors shaping codon choices of single stranded DNA (ssDNA) and double stranded DNA (dsDNA) Escherichia coli phages. In chapter two, we employed two indices, rRSCU (correlation in relative synonymous codon usage between phages and their hosts) and CAI (codon adaptation index) to measure codon adaptation in phages. None of the analyzed ssDNA phages encode tRNAs while some dsDNA phages encode their own tRNAs. Both rRSCU and CAI are negatively correlated with number of tRNA genes encoded by these dsDNA phages. We observed significantly greater rRSCU for dsDNA phages (without tRNAs) than ssDNA phages. In addition, we propose that ssDNA phages have evolved a novel codon adaptation strategy to overcome the disruptive effect of their high C→T mutation rates in codon adaptation with host. In chapter three, we formulated an index phi to measure selection by host translation machinery and to present explicit linear and nonlinear models to characterize the effect of C→T mutation and host-tRNA-mediated selection on phage codon usage. The effect of selection (phi) on codon usage is detectable in most dsDNA and ssDNA phage species. C→T mutations also interfere with nonsynonymous substitutions at second codon positions, especially in ssDNA phages. Strand asymmetry along with the accompanying local variation in mutation bias can significantly affect codon adaptation in both dsDNA and ssDNA phages.

Bacteriophages (hereafter referred as phages) can translate their mRNAs efficiently by maximizing the use of codons decoded by the most abundant tRNAs of their bacterial hosts. Translation efficiency directly influences phage fitness and evolution. Reengineered phages find application in controlling their host population in both health and industry. The objective of this thesis work is to examine the factors shaping codon choices of single stranded DNA (ssDNA) and double stranded DNA (dsDNA) Escherichia coli phages. In chapter two, we employed two indices, rRSCU (correlation in relative synonymous codon usage between phages and their hosts) and CAI (codon adaptation index) to measure codon adaptation in phages. None of the analyzed ssDNA phages encode tRNAs while some dsDNA phages encode their own tRNAs. Both rRSCU and CAI are negatively correlated with number of tRNA genes encoded by these dsDNA phages. We observed significantly greater rRSCU for dsDNA phages (without tRNAs) than ssDNA phages. In addition, we propose that ssDNA phages have evolved a novel codon adaptation strategy to overcome the disruptive effect of their high C→T mutation rates in codon adaptation with host. In chapter three, we formulated an index phi to measure selection by host translation machinery and to present explicit linear and nonlinear models to characterize the effect of C→T mutation and host-tRNA-mediated selection on phage codon usage. The effect of selection (phi) on codon usage is detectable in most dsDNA and ssDNA phage species. C→T mutations also interfere with nonsynonymous substitutions at second codon positions, especially in ssDNA phages. Strand asymmetry along with the accompanying local variation in mutation bias can significantly affect codon adaptation in both dsDNA and ssDNA phages.

Primates are one of the best characterized phylogenies with vast amounts of comparative data available, including genomic sequences, gene expression, and epigenetic modifications. Thus, they provide an ideal system to study sequence evolution, regulatory evolution, epigenetic evolution as well as their interplays. Comparative studies of primate genomes can also shed light on molecular basis of human-specific traits. This dissertation is mainly composed of three chapters studying human and non-human primate evolution. The first study investigated evolutionary rate difference between sex chromosome and autosomes across diverse primate species. The second study developed an unbiased approach without the need of prior information to identify genomic segments under accelerated evolution. The third study investigated interplay between genomic and epigenomic evolution of humans and chimpanzees. Research advance 1: evolutionary rates of the X chromosome are predicted to be different from those of autosomes. A theory based on neutral mutation predicts that the X chromosome evolves slower than autosomes (slow-X evolution) because the numbers of cell division differ between spermatogenesis and oogenesis. A theory based on natural selection predicts an opposite direction (fast-X evolution) because newly arising beneficial mutations on the autosomes are usually recessive or partially recessive and not exposed to natural selection. A strong slow-X evolution is also predicted to counteract the effect of fast-X evolution. In our research, we simultaneously studied slow-X evolution, fast-X evolution as well as their interaction in a phylogeny of diverse primates. We showed that slow-X evolution exists in all the examined species, although their degrees differ, possibly due to their different life history traits such as generation times. We showed that fast-X evolution is lineage-specific and provided evidences that fast-X evolution is more evident in species with relatively weak slow-X evolution. We discussed potential contribution of various degrees of slow-X evolution on the conflicting population genetic inferences about human demography. Research advance 2: human-specific traits have long been considered to reside in the genome. There has been a surge of interest to identify genomic regions with accelerated evolution rate in the human genome. However, these studies either rely on a priori knowledge or sliding windows of arbitrary sizes. My research provided an unbiased approach based on previously developed “maximal segment” algorithm to identify genomic segments with accelerated lineage-specific substitution rate. Under this framework, we identified a large number of human genomic segments with clustered human-specific substitutions (named “maximal segments” after the algorithm). Our identified human maximal segments cover a significant amount of previously identified human accelerated regions and overlap with genes enriched in developmental processes. We demonstrated that the underlying evolutionary forces driving the maximal segments included regionally increased mutation rate, biased gene conversion and positive selection. Research advance 3: DNA methylation is one of the most common epigenetic modifications and plays a significant role in gene regulation. How DNA methylation status varies on the evolutionary timescale is not well understood. In this study, we investigated the role of genetic changes in shaping DNA methylation divergence between humans and chimpanzees in their sperm and brain, separately. We find that for orthologous promoter regions, CpG dinucleotide content difference is negatively correlated with DNA methylation level difference in the sperm but not in the brain, which may be explained by the fact that CpG depleting mutations better reflect germline DNA methylation levels. For the aligned sites of orthologous promoter regions, sequence divergence is positively correlated with methylation divergence for both tissues. We showed that the evolution of DNA methylation can be affected by various genetic factors including transposable element insertions, CpG depleting mutations and CpG generating mutations.