How to read a phylogenetic tree

A phylogenetic tree is a hypothesis about evolutionary history. It shows which groups share more recent common ancestors than others, and it gets revised when new evidence, especially DNA sequence data, comes in. Reading one correctly comes down to a single rule: relatedness is set by how recently two lineages share an ancestor, not by how close they sit on the page. Everything below is a way of applying that rule.

An example tree

Take a small rooted tree for five vertebrates. Written in nested form, where each pair of parentheses is one branch point, it is:

(lamprey, (perch, (lizard, (rabbit, human))))

Drawn out, the root is on the left and the lineages split one at a time as you move right. These are the four branch points, from oldest to most recent:

Sketch it on paper before reading on. Every question below uses this tree.

The vocabulary

• Root: the common ancestor of everything on the tree. A rooted tree has one; an unrooted tree shows relationships among the included groups without saying which lineage diverged first.
• Branch point (node): where one lineage splits into two or more. Each node stands for the most recent common ancestor of everything to its right.
• Sister taxa: two lineages that split from the same node and nothing else. Rabbit and human are sister taxa (node 3).
• Basal taxon: a lineage that branched off early and shows little branching of its own. Here the lamprey is basal.
• Polytomy: a node with three or more branches coming out of it. It means the branching order among those lineages isn't resolved yet, not that they all split at the same instant.
• Clade: a common ancestor plus all of its descendants, also called a monophyletic group. Lizard, rabbit and human form a clade (everything from node 2). Perch plus lizard, leaving out rabbit and human, is not a clade.

Question 1: is the human more closely related to the rabbit or the lizard?

The rabbit. Human and rabbit share node 3, which is more recent than node 2, the ancestor they each share with the lizard.

Question 2: is the lizard more closely related to the rabbit or the human?

Neither. The lizard joins the rabbit-human lineage at node 2, so it is equally related to both. This is the question students most often get wrong, because in most drawings the lizard sits right next to the rabbit. Position along the tips means nothing.

Rotating branches changes nothing

You can swivel any node like a mobile. Flip node 3 and the tips read human, rabbit instead of rabbit, human; flip node 2 and the lizard moves from above the pair to below it. The tree says exactly the same thing, because the branching pattern (the topology) is unchanged. On an exam, if two drawings look different, trace the nodes before deciding they disagree.

A related trap is reading the tips as a ladder of progress. The lamprey is not a "less evolved" human. Every living lineage on the tree has been evolving for exactly the same amount of time since the root.

Branch length usually means nothing

Unless the tree is labeled with a time scale or a scale bar for amount of change, branch length carries no information. Only the branching pattern does. If your sketch gives the lamprey a long branch because it splits off first, that length is just a drawing choice.

Characters: ancestral versus derived

Trees are built from characters, traits that are present in some groups and absent in others. Map these onto the example:

• A vertebral column (or vertebral elements, in the lamprey): all five have one
• Hinged jaws: perch, lizard, rabbit, human (not the jawless lamprey)
• Amniotic egg: lizard, rabbit, human
• Hair: rabbit and human

A shared ancestral character is present in the common ancestor of the whole group being compared, so it can't tell members apart. The backbone is shared ancestral for this tree. A shared derived character evolved later, inside the group, and so marks a clade. The amniotic egg is shared derived: it arose on the branch leading to node 2, and it groups the lizard, rabbit and human while leaving out the perch. Hair does the same for node 3.

Whether a trait is ancestral or derived depends on the comparison. For a tree of only mammals, hair would be shared ancestral, because every mammal has it.

How parsimony picks a tree

Maximum parsimony says to prefer the tree that needs the fewest evolutionary changes to explain the traits. On the example tree, each of the three traits that vary (hinged jaws, amniotic egg, hair) evolves exactly once, so it needs 3 changes.

Now try an alternative tree where the lizard and human are sister taxa and the rabbit branches off before them. Hair now has to be explained either by evolving twice (once in the rabbit lineage, once in the human lineage) or by evolving once and then being lost in the lizard. Either way that is 4 changes instead of 3. Parsimony prefers the original tree.

Homologous versus analogous traits

Only homologous traits, ones inherited from a common ancestor, are evidence of relationship. Bat wings and bird wings are homologous as forelimbs: the same underlying bones, modified for flight in each lineage. Insect wings and bird wings are analogous: they do the same job but evolved independently, through convergent evolution. Grouping organisms by an analogous trait would build a wrong tree, which is why molecular data is so useful as a check.

Where trees break down

Horizontal gene transfer moves genes between unrelated lineages, sideways across branches, and it is common in prokaryotes (through transformation, transduction and conjugation). A strictly branching tree can't show that, which is why some biologists represent early life as a web or a ring rather than a single trunk. For the course, read trees as described above, and know that HGT is the main reason the base of the tree of life is uncertain. When a tree is drawn to a time scale, the dates usually come from fossils dated by radioactive decay, which is the same first-order half-life math you meet in chemistry.

Practice

Phylogenetics is its own unit in General Biology II: what's in General Biology II shows where it sits. The unit right before it covers Hardy-Weinberg equilibrium, and how to study for General Biology II covers drawing trees from memory as practice. Tree reading is a skill you keep only by doing it, so space your practice out; what spaced repetition actually does explains why.

Encodr's free General Biology II course includes a full chapter on reading phylogenetic trees, plus cladistics and horizontal gene transfer.