Molecular Evolution

Molecular evolution is the branch of evolutionary biology and genomics that investigates how genetic information encoded in deoxyribonucleic acid (DNA), ribonucleic acid (RNA), and protein sequences changes across generations and geological time. Founded in the 1960s by Emile Zuckerkandl and Linus Pauling (who introduced the molecular clock hypothesis in 1962–1965) and given its quantitative population-genetic foundation by Motoo Kimura (1968), Jack Lester King, Thomas Jukes (1969), and Tomoko Ohta (1973), molecular evolution integrates biochemical mutation mechanics—point substitutions, insertions/deletions (indels), gene duplications, and horizontal gene transfer—with the population dynamics of random genetic drift, purifying selection, and positive selection.

Within comparative genomics and phylogenetic cladistics, molecular evolution provides the stochastic Markov models required to translate sequence alignments into evolutionary distances and divergence times. By distinguishing neutral polymorphisms that drift at the baseline mutation rate from functionally constrained sites and adaptive substitutions (evaluated via codon substitution models and the Felsenstein pruning algorithm), molecular evolution reconstructs the genealogical history of genes, genomes, and species under the multispecies coalescent.

Kimura's Neutral Fixation Identity (k = μ₀) and the Molecular Clock

Prior to protein sequencing and allozyme electrophoresis in the 1960s, classical neo-Darwinian theory assumed that the vast majority of amino acid and nucleotide substitutions fixed between species were driven by Darwinian positive selection. In 1968, Motoo Kimura calculated from mammalian hemoglobin, cytochrome c, and fibrino-peptide divergence rates that the genome-wide nucleotide substitution rate in mammals would impose an unsustainable mutational genetic load (Haldane's dilemma) if every substitution were selective. Kimura therefore proposed the Neutral Theory of Molecular Evolution, positing that the majority of evolutionary changes at the molecular level—and most within-species genetic polymorphisms—result not from positive Darwinian selection on advantageous alleles, but from the random fixation of selectively neutral or nearly neutral mutants through random genetic drift.

In a diploid population of effective size N_e experiencing a neutral mutation rate μ₀ per site per generation, a total of 2*N*_e μ₀ new neutral mutations arise across the population each generation. Because each newly arisen neutral allele begins at an initial frequency of p₀ = 1 / (2*N*_e) and has a fixation probability exactly equal to its initial frequency (u = 1 / (2*N*_e)), the long-term rate of neutral substitution k per unit time equals the product of the population mutation input and fixation probability:

This cancellation of effective population size N_e yields Kimura's foundational result: for strictly neutral mutations, the rate of evolutionary substitution (k) equals the neutral mutation rate (μ₀), independent of population size. Provided μ₀ remains approximately constant per year or per generation across lineages, sequence divergence accumulates linearly with time (D ≈ 2*μ*₀t), establishing the mechanistic basis of Zuckerkandl and Pauling's molecular clock.

Tomoko Ohta's Nearly Neutral Theory (|N_e s| ≈ 1) and Purifying Constraint

In 1973, Tomoko Ohta extended Kimura's framework into the Nearly Neutral Theory of Molecular Evolution by analyzing mutations sitting at the boundary between genetic drift and natural selection, where the scaled selection coefficient satisfies |*N*_e *s*| ≈ 1 (or |4*N*_e *s*| ≤ 4). Under Kimura's diffusion approximation for a mutation with selection coefficient s in a diploid population:

Slightly deleterious mutations (s < 0, with N_e|*s*| ≈ 0.2 to 2) are purged efficiently by purifying selection in species with large effective population sizes (N_e > 10⁶, such as pelagic bacteria or Drosophila), but behave as effectively neutral and drift to fixation in species with small effective population sizes (N_e ≈ 10⁴, such as large-bodied vertebrates). Because generation time g typically scales inversely with body size while N_e scales negatively with body size, Ohta's nearly neutral model accounts for why protein clocks often tick more uniformly with chronological time (years) than with generation count. Across functional genomes, the fraction f₀ of mutations spared by purifying selection determines the relative evolutionary rate: fibrinopeptides and pseudogenes (f₀ ≈ 1.0) evolve near the unconstrained mutation rate (~4–9 × 10⁻⁹ substitutions site⁻¹ yr⁻¹ in mammals), hemoglobins evolve at an intermediate rate (~1 × 10⁻⁹ site⁻¹ yr⁻¹), and core histone H4 and ubiquitin (f₀ < 0.01) accumulate fewer than 10⁻¹¹ amino acid replacements site⁻¹ yr⁻¹.

Transition/Transversion Bias (κ), GTR Matrices, and Gamma Rate Heterogeneity (+Γ)

When two homologous DNA sequences diverge from a common ancestor over time t, the observed proportion of differing sites (p-distance) underestimates the true number of evolutionary substitutions (d = 2*μt*) because multiple substitutions ("multiple hits"), parallel changes, and back-mutations occur at the same nucleotide position—saturating toward p = 0.75 under uniform base composition as t → ∞. Molecular evolution models sequence divergence as a continuous-time Markov chain (CTMC) specified by a 4 × 4 instantaneous rate matrix Q = {*q*_ij}, yielding transition probabilities P(t) = exp(Qt).

Whereas the Jukes–Cantor (JC69) model assumes all nucleotide substitutions occur at an identical rate, empirical DNA alignments exhibit a two- to ten-fold excess of transitions (purine↔purine A↔G or pyrimidine↔pyrimidine C↔T, propelled by spontaneous cytosine deamination of 5-methylcytosine at CpG dinucleotides and tautomeric base pairing) over transversions (purine↔pyrimidine). Kimura's two-parameter (K80 / K2P) model corrects for this transition/transversion ratio κ = α/β, while the General Time Reversible (GTR) model allows six distinct symmetric exchangeability rates paired with four equilibrium base frequencies (π_A, π_C, π_G, π_T). Because substitution rates vary substantially across sites (e.g., rapid 3rd codon positions vs. invariant catalytic 2nd codon positions, or rRNA loops vs. stems), Ziheng Yang (1994) introduced discrete gamma-distributed among-site rate variation (+Γ₄) with shape parameter α, which models skewed substitution rates when evaluating tree likelihoods via the Felsenstein pruning algorithm.

Ohno's Neofunctionalization, DDC Subfunctionalization, and Concerted Evolution

Beyond point substitution, the primary source of new genes and biochemical pathways is gene and genome duplication, emphasized by Susumu Ohno in Evolution by Gene Duplication (1970). Duplicate genes produced by unequal crossing-over, retrotransposition, or whole-genome polyploidization (such as the two rounds of 2R whole-genome duplication at the base of vertebrates) experience four evolutionary fates:

1. Nonfunctionalization (Pseudogenization): The most frequent fate; one copy accumulates disabling frameshift or nonsense mutations and decays into a neutral pseudogene within a few million years.
2. Neofunctionalization: While one paralog remains under purifying selection to maintain the ancestral physiological function, the redundant sister copy accumulates mutations that occasionally confer an entirely novel adaptive activity under positive selection (e.g., digestive ribonucleases and antifreeze glycoproteins).
3. Subfunctionalization (Duplication–Degeneration–Complementation, DDC): Formulated by Michael Lynch and John Force (1999), complementary degenerative mutations in modular cis-regulatory enhancers partition the ancestral expression profile between the two daughter paralogs, rendering both copies essential for organismal survival without requiring positive selection.
4. Gene Conversion and Concerted Evolution: In tandem multigene arrays (such as eukaryotic 18S–5.8S–28S ribosomal DNA repeats), homologous recombination and non-reciprocal gene conversion homogenize paralogous copies within a species so they evolve in concert rather than diverging independently.

Does Kimura's Neutral Theory claim that most mutations are neutral and that natural selection is unimportant?

No. Kimura's Neutral Theory explicitly recognizes that the majority of newly arising non-synonymous mutations in functional coding regions are deleterious and are rapidly eliminated by purifying (negative) natural selection. What the Neutral Theory states is that among the mutations that survive purifying selection to reach fixation as inter-species substitutions (or persist as polymorphic alleles within populations), the vast majority are selectively neutral or nearly neutral rather than driven by positive Darwinian selection.

Why do third codon positions evolve much faster than second codon positions?

Due to the degeneracy of the standard genetic code, roughly 70% of nucleotide changes at the third codon position (especially 4-fold degenerate sites such as GGN for glycine or ACN for threonine) are synonymous—they do not alter the encoded amino acid sequence and therefore escape protein-level purifying selection. Conversely, 100% of nucleotide changes at the second codon position are non-synonymous, altering the chemical polarity or size of the encoded amino acid and facing strong purifying selection.

Why is multiple-hit correction necessary when estimating evolutionary distances from DNA sequences?

When two sequences have diverged recently (p < 0.05 differences per site), the observed fraction of mismatched nucleotides (p-distance) closely approximates the true number of substitutions (d). However, as divergence time increases, new mutations repeatedly strike sites that have already mutated—including parallel substitutions and reversals back to the ancestral base. Without applying a Markov substitution model (such as K2P, HKY85, or GTR+Γ), saturation causes observed p-distances to plateau below 0.75 and systematically underestimates deep branch lengths.

1. Zuckerkandl, E., & Pauling, L. (1965). "Evolutionary divergence and convergence in proteins." In V. Bryson & H. J. Vogel (Eds.), Evolving Genes and Proteins (pp. 97–166). Academic Press. https://doi.org/10.1016/B978-1-4832-2734-4.50017-6
2. Kimura, M. (1968). "Evolutionary rate at the molecular level." Nature, 217(5129), 624–626. https://doi.org/10.1038/217624a0
3. Ohta, T. (1973). "Slightly deleterious mutant substitutions in evolution." Nature, 246(5428), 96–98. https://doi.org/10.1038/246096a0
4. Kimura, M. (1980). "A simple method for estimating evolutionary rates of base substitutions through comparative studies of nucleotide sequences." Journal of Molecular Evolution, 16(2), 111–120. https://doi.org/10.1007/BF01731581