How Phylogenetics Works

Scientists extract DNA, RNA, or protein sequences from different organisms and convert them into long strings of letters representing nucleotides or amino acids. For DNA, these are the familiar A, T, G, and C bases that form the genetic code. A human hemoglobin gene might be compared with the corresponding gene in chimpanzees, gorillas, and more distant relatives. Modern sequencing technology can generate millions of these letters quickly and cheaply, providing raw material for evolutionary analysis.

The key insight is that closely related species accumulate fewer differences in their sequences than distantly related ones. If humans and chimpanzees share 98.8% of their DNA sequence while humans and mice share only 85%, this numerical difference reflects evolutionary time since divergence. Each sequence acts like a molecular document preserving the history of mutations that occurred over millions of years. Scientists typically focus on specific genes or genomic regions that evolve at useful rates—not too fast to erase the signal, not too slow to show meaningful variation.

Phylogenetic comparison involves identifying homologous features—traits inherited from a common ancestor rather than independently evolved. A bat's wing, a human's arm, and a whale's flipper all contain the same arrangement of bones (humerus, radius, ulna) despite serving different functions, indicating these mammals share a common ancestor with that skeletal pattern. Scientists distinguish these homologies from analogies, where similar features evolved independently, like the wings of bats and insects.

For molecular data, researchers align sequences to match up corresponding positions across different species. If a particular DNA site contains an "A" in humans, chimpanzees, and gorillas but a "G" in monkeys, this shared "A" suggests the great apes inherited it from their exclusive common ancestor. Computer algorithms slide sequences against each other, inserting gaps where necessary, to maximize matches and identify these informative positions.

The comparison process also weights different types of changes. Some amino acid substitutions dramatically alter protein function while others are biochemically similar and more easily tolerated. Likewise, some DNA positions are under strong natural selection and change slowly, while others in non-coding regions accumulate mutations freely. By carefully comparing both what has changed and what has remained constant, scientists extract meaningful evolutionary signals from the noise of random variation.

Once sequences or traits are compared, phylogenetics applies mathematical models to quantify how much evolutionary change separates different organisms. A simple count might show that two DNA sequences differ at 15 out of 1,000 positions, yielding 1.5% divergence. However, this raw percentage underestimates true evolutionary distance because some sites may have mutated multiple times, with later changes erasing evidence of earlier ones. If a position changed from A to G, then later from G back to A, a simple comparison would miss both events.

Sophisticated models correct for this hidden complexity using probability theory and assumptions about how evolution proceeds. The Jukes-Cantor model assumes all nucleotide substitutions occur at equal rates, while more realistic models like GTR allow different rates for transitions (A↔G, C↔T) versus transversions (A↔C, A↔T, G↔C, G↔T). These models use calculus to estimate the actual number of changes that occurred, not just the differences we observe today.

The resulting distance matrix creates a mathematical landscape where each species occupies a position, and the distances between them reflect evolutionary time or genetic change. Two species separated by a large calculated distance have experienced more mutations or longer independent evolution than those with a small distance. These quantified relationships become the foundation for reconstructing the actual tree structure showing how species connect.

Reconstructing a phylogenetic tree means determining the branching pattern that best explains the observed data. For just ten species, there are over 34 million possible tree shapes, and the number grows astronomically with more organisms. Scientists use computational algorithms to search this vast space efficiently. Distance-based methods like neighbor-joining start by clustering the most similar species, then progressively join more distant relatives. Character-based methods like maximum parsimony seek trees requiring the fewest evolutionary changes, while maximum likelihood finds trees that make the observed data most probable given a specific evolutionary model.

Each tree consists of branches representing lineages and nodes representing common ancestors where lineages split. Branch lengths often indicate evolutionary distance—longer branches mean more accumulated changes. Internal nodes aren't labeled with specific ancestor names because we rarely know exactly what those extinct organisms were, but their positions show when different groups diverged. A tree might show that rodents and primates split 90 million years ago at one node, while humans and chimpanzees split 7 million years ago at a much more recent node.

Modern phylogenetics also quantifies confidence in tree topology using statistical techniques like bootstrapping, which repeatedly resamples the data to see if the same branching pattern emerges consistently. A branch supported by 95% of bootstrap replicates is considered highly reliable, while one below 70% support is questionable. This statistical rigor distinguishes robust evolutionary relationships from ambiguous ones where data is insufficient or conflicting.

Tracing lineage means reading the phylogenetic tree from tips (present-day species) back through time to increasingly ancient common ancestors. Every time you encounter a node moving rootward, you're traveling backward to when a single ancestral population split into separate evolutionary paths. Following humans back through time, we trace to a node shared with chimpanzees around 7 million years ago, then further back to a node shared with gorillas at roughly 10 million years, then to Old World monkeys at 25 million years, and eventually to all mammals beyond 200 million years.

This temporal framework lets scientists map when key evolutionary innovations appeared. By examining which species possess specific traits and where those traits appear on the tree, researchers can infer that feathers evolved in the lineage leading to birds after it split from crocodiles, or that lactation evolved in the mammalian stem lineage before the diversification of marsupials and placentals. Fossils provide calibration points, allowing scientists to convert branch lengths from genetic units into actual years by matching extinct species to their phylogenetic positions.

Lineage tracing also reveals patterns like adaptive radiation, where a single ancestral species rapidly diversifies into many forms occupying different ecological niches. Darwin's finches on the Galápagos Islands trace back to a single colonizing ancestor, with subsequent branches showing how different beak shapes evolved for processing different foods. By following these lineages, phylogenetics transforms evolution from an abstract concept into a concrete historical sequence with identifiable stages, timings, and transformations.