Loading field map
How do genomes evolve, and how can we detect the forces—mutation, drift, selection, recombination—that shape them? This question has driven evolutionary genomics since its emergence as a distinct subfield in the late 1960s. The frameworks that have defined the field are not simply a list of discoveries; they are competing and complementary systems of assumptions, methods, and explanatory commitments. Each framework changed what counted as a good explanation for genomic patterns, and their relationships—refinement, infrastructure, supplementation, and sometimes productive tension—form the real story of the field.
Before 1968, most evolutionary biologists assumed that natural selection was the dominant force at every level, including the molecular one. This pan-selectionist view left little room for neutral or nearly neutral changes. Motoo Kimura’s Neutral Theory of Molecular Evolution (1968) challenged that assumption head-on. Kimura argued that the vast majority of molecular differences between species are caused by mutation and genetic drift, not by selection. The theory did not deny that selection acts on some sites; it provided a null model against which selection could be detected. If the observed pattern of molecular variation matches the neutral expectation, there is no reason to invoke selection. If it deviates, selection becomes a candidate explanation. This shift was foundational: it gave evolutionary genomics its central statistical tool—the neutral null model—and its central puzzle: distinguishing the genomic signatures of drift from those of selection.
Almost immediately, the Neutral Theory faced a challenge from within. Tomoko Ohta’s Nearly Neutral Theory (1973) argued that the sharp boundary between strictly neutral and strongly selected mutations was unrealistic. Many mutations have selection coefficients too small for natural selection to act efficiently, especially in small populations. Their fate is governed by a mixture of drift and weak selection, and their behavior depends on population size. Ohta’s framework did not replace Kimura’s; it refined it by introducing a continuum. The Nearly Neutral Theory preserved the neutral null model as a starting point but added a population-size-dependent layer that explained patterns the original theory could not, such as the correlation between genome size and effective population size. Today, most empirical tests of selection use a nearly neutral baseline rather than a strictly neutral one.
A major methodological breakthrough came with Coalescent Theory (1982), developed by John Kingman and others. Earlier population genetics modeled populations forward in time, tracking allele frequencies generation by generation. The coalescent turned this around: it traces the genealogy of a sample of sequences backward to their most recent common ancestor. This backward perspective dramatically simplified calculations and made it possible to infer population history—effective population size, migration rates, growth, and divergence times—directly from genetic data. The coalescent provided the statistical infrastructure that the neutral and nearly neutral frameworks needed. It allowed researchers to generate explicit null distributions for summary statistics (such as Tajima’s D) and to test whether observed genomic patterns fit neutral expectations. Without the coalescent, the neutral null model would have remained a theoretical idea rather than a practical tool for genomic analysis.
By the mid-1990s, genome sequencing had advanced enough to make whole-genome comparisons feasible. Comparative Genomics (1995) shifted the unit of analysis from individual genes to entire genomes. The framework’s core method is alignment: aligning genomes from different species to identify conserved regions (likely functional) and divergent regions (likely evolving neutrally or under positive selection). Comparative genomics built directly on the neutral null model: conserved elements are those that evolve slower than the neutral rate, implying purifying selection; rapidly evolving elements suggest positive selection or relaxed constraint. The framework also revealed unexpected phenomena, such as the extensive conservation of non-coding regulatory regions and the widespread presence of transposable elements. Comparative genomics did not replace the earlier frameworks; it scaled them up, turning the neutral model into a genome-wide search tool.
Around 2000, a group of evolutionary biologists began arguing that the Modern Evolutionary Synthesis—the mid-20th-century integration of Darwinian selection with Mendelian genetics—needed updating. The Extended Evolutionary Synthesis (EES) (2000) proposed that evolutionary change is not solely driven by natural selection acting on random genetic variation. It emphasized developmental plasticity, niche construction, inclusive inheritance (including epigenetic and behavioral inheritance), and the role of organisms as co-constructors of their environments. For evolutionary genomics, the EES raised new questions: How much of genomic variation is shaped by non-genetic inheritance? Do epigenetic marks evolve under selection? Can developmental bias channel the direction of genomic change? The EES did not reject the neutral or nearly neutral frameworks; it supplemented them by broadening the range of heritable variation and evolutionary processes that genomic data might reveal. Its influence on genomics remains modest but growing, particularly in studies of epigenetic variation and gene regulation.
Phylogenomics (2000) emerged at roughly the same time as the EES, but from a different lineage: the intersection of phylogenetics and genomics. Traditional phylogenetics reconstructed evolutionary trees from a handful of genes. Phylogenomics uses genome-scale data—hundreds or thousands of genes, or whole genomes—to infer species trees, resolve deep evolutionary relationships, and study the evolution of gene families. The framework introduced the critical distinction between orthologs (genes related by speciation) and paralogs (genes related by duplication), which is essential for accurate tree-building. Phylogenomics also revealed widespread gene tree discordance, caused by incomplete lineage sorting, horizontal gene transfer, and gene duplication and loss. This complexity forced the development of new methods that account for the fact that different parts of a genome can have different evolutionary histories. Phylogenomics extended comparative genomics by adding a rigorous phylogenetic framework, and it coexists with coalescent-based methods for population-level inference.
The most recent major framework, Evolutionary Systems Biology (2005), represents a shift from studying individual genes or proteins to studying the evolution of molecular networks—gene regulatory networks, metabolic pathways, protein-protein interaction networks. The framework asks how network structure evolves, whether network properties (such as modularity or robustness) are shaped by selection, and how changes in network architecture produce phenotypic novelty. Evolutionary systems biology draws on both comparative genomics (for identifying conserved network components) and the EES (for thinking about how developmental and environmental factors shape network evolution). It also uses mathematical modeling and high-throughput data to test hypotheses about network-level selection. The framework does not replace earlier approaches; it adds a new level of analysis—the system—that cannot be reduced to the sum of its parts.
Today, no single framework dominates evolutionary genomics. The neutral and nearly neutral models remain the default null for detecting selection, and coalescent theory provides the statistical machinery for most population-genomic analyses. Comparative genomics and phylogenomics are the standard tools for studying genome function and evolutionary history across deep timescales. The EES and evolutionary systems biology are more recent and less universally adopted, but they are gaining traction as genomic data reveal phenomena—epigenetic variation, network-level conservation, developmental plasticity—that earlier frameworks did not foreground.
The main active disagreements center on how much of genomic evolution is driven by selection versus drift (a debate that traces back to the Neutral Theory’s challenge to pan-selectionism), whether non-genetic inheritance plays a significant evolutionary role (the EES versus more gene-centric views), and whether network-level properties are targets of selection or byproducts of lower-level processes. These disagreements are not signs of weakness; they reflect the field’s maturity. Evolutionary genomics is a pluralistic discipline in which different frameworks are suited to different questions. The neutral model is indispensable for detecting selection; the coalescent is indispensable for inferring population history; phylogenomics is indispensable for reconstructing deep evolutionary relationships; and the EES and systems biology are indispensable for understanding the full complexity of how genomes evolve. The field’s progress depends on keeping these frameworks in productive tension rather than forcing a premature unification.