← Back to course

Phylogenetics & Cladistics

You’ll be able to

Only shared derived characters build clades

Cladistics groups organisms strictly by shared derived characters — evolutionary novelties, called synapomorphies, that arose in a common ancestor and were inherited by all its descendants. A synapomorphy at one level can be an ancestral (plesiomorphic) trait at a deeper level: a backbone is derived for vertebrates but ancestral within mammals. The key rule is that only synapomorphies define a clade; a shared ancestral character (symplesiomorphy) does not, because it was already present before the group arose and so cannot distinguish its members from outsiders. A clade (monophyletic group) is an ancestor plus all of its descendants — a group you could snip off the tree with a single cut.

Outgroups and parsimony do the actual work

To decide which state of a character is ancestral versus derived, cladists include an outgroup — a taxon known to have branched off before the group of interest (the ingroup). Whatever state the outgroup shows is taken as the ancestral condition, so any different state within the ingroup is derived and potentially informative. With characters polarized this way, the preferred tree is chosen by maximum parsimony: of all possible trees, favor the one that requires the fewest evolutionary changes. Extra changes usually mean homoplasy — the same trait arising independently more than once (convergence) or being lost — which parsimony treats as a more costly, less likely explanation.

Molecular clocks and clashing signals

Because many DNA and protein sequences accumulate mutations at a roughly steady average rate, the number of differences between two species estimates how long ago they diverged — a molecular clock. The clock must be calibrated against an independently dated event (often a fossil) to convert percent sequence difference into years. Molecular data sometimes conflict with morphology: convergent evolution can make unrelated species look alike (homoplasy), so a tree built from body shape may group them wrongly, while sequence data — with far more independent characters — usually resolves the true branching order. When morphological and molecular trees disagree, biologists weigh which characters are more likely to be homoplastic.

Maximum parsimony
preferred tree = the tree requiring the fewest character-state changes
Fewer changes means less assumed homoplasy (convergence or reversal). Among competing trees, the simplest explanation is favored.
Molecular clock
sequence divergence ≈ rate × time → time = divergence ÷ rate
Calibrate the rate using a node whose age is known from the fossil record, then apply it to date other divergences.
Worked example

You are given this character matrix (1 = trait present, 0 = absent) with an outgroup: Outgroup 0-0-0, Fish 1-0-0, Frog 1-1-0, Mouse 1-1-1, for the characters (vertebral column, four limbs, hair). Build the most parsimonious cladogram and state where each character arose.

  1. 1.Use the outgroup to polarize: it has none of the three traits, so absence (0) is ancestral and presence (1) is derived for each character.
  2. 2.Order taxa by how many derived traits they share: Fish has the vertebral column only; Frog adds four limbs; Mouse adds hair — a nested set.
  3. 3.Place a vertebral column at the base of the ingroup (uniting Fish + Frog + Mouse), then four limbs on the branch to Frog + Mouse, then hair on the branch to Mouse.
  4. 4.Count the changes: each of the three characters is gained exactly once — 3 changes total, with no trait appearing or disappearing twice.
  5. 5.Confirm parsimony: any tree that, for example, grouped Fish with Mouse would force a trait like four limbs or hair to evolve twice (homoplasy), requiring more than 3 changes — so it is rejected.
Answer: The most parsimonious tree is a nested ladder: (Outgroup, (Fish, (Frog, Mouse))). The vertebral column is a synapomorphy uniting all vertebrates, four limbs unite Frog + Mouse, and hair is unique to Mouse — three character changes total, the fewest possible.
Worked example

A protein is used as a molecular clock. Fossils show that species A and B diverged 20 million years ago, and their sequences differ by 4%. Species A and C differ by 10% in the same protein. Estimate when A and C diverged.

  1. 1.Calibrate the clock with the fossil-dated node: A and B differ by 4% over 20 million years, so the rate = 4% ÷ 20 My = 0.2% per million years.
  2. 2.Assume the same rate applies to the A–C comparison (a constant clock).
  3. 3.Apply time = divergence ÷ rate: time = 10% ÷ (0.2% per My) = 50 million years.
  4. 4.Interpret: the larger sequence difference (10% vs 4%) correctly implies a deeper, older split.
Answer: Species A and C are estimated to have diverged about 50 million years ago (10% ÷ 0.2% per My = 50 My), using the rate of 0.2% per million years calibrated from the fossil-dated A–B split.
Checkpoint

Which of the following describes a monophyletic group (a valid clade)?

Checkpoint

Two competing trees explain the same character data. Tree 1 requires 6 evolutionary changes; Tree 2 requires 9. By the principle of maximum parsimony, which tree is preferred and why?

Watch out

Do not confuse a synapomorphy (shared derived character — the only thing that defines a clade) with a symplesiomorphy (shared ancestral character, which does not). And do not read relatedness from how physically alike two species look: convergent evolution produces homoplasy, so similar-looking organisms can sit far apart on the true tree.

Checkpoint

A cladogram built only from wing shape groups bats together with birds, but DNA sequence data place bats firmly among the other mammals. What is the best explanation?

On the exam

For tree questions, work in this order: use the outgroup to decide which states are derived, group taxa by shared derived characters (synapomorphies), and pick the tree with the fewest changes (parsimony). Treat a morphology-versus-molecule conflict as a homoplasy problem — name convergent evolution explicitly and explain why the character misleads.

Answer the 3 checkpoints as you read.

Sign in to save your progress