Questions: Gene Tree and Species Tree Incongruence: Lineage Sorting

Question 1

You sequence 1,000 genes from three closely related bird species (A, B, C). The species tree places A and B as sister species. You find 420 genes recover (A,B), 310 genes recover (A,C), and 270 genes recover (B,C). How should you interpret this result?

• A. The true species tree is uncertain — you cannot trust any topology when only 42% of genes agree
• B. The roughly equal frequencies of the two discordant topologies are consistent with incomplete lineage sorting: random ancestral sorting predicts the two alternative discordant trees should occur at approximately equal rates
• C. The discordant trees indicate hybridization between species A–C and between B–C, because multiple topologies are recovered
• D. The majority topology (A,B) is the species tree; the discordant trees represent sequencing errors that should be filtered out

Answer: (B)

ILS from random lineage sorting predicts a specific pattern: when discordance occurs, the two alternative discordant topologies (A,C) and (B,C) should appear at roughly equal frequencies, because there is no directional force favoring one over the other — ancestral alleles simply sort randomly. Here (310 vs 270) is roughly equal. Hybridization, by contrast, would produce directional discordance: one alternative topology would be enriched relative to the other because gene flow has a direction. Options A and D misunderstand that discordance under ILS is expected and informative, not erroneous.

Question 2

Why does rapid adaptive radiation dramatically increase the probability of gene-tree and species-tree incongruence?

• A. Rapid speciation increases mutation rates, generating more homoplasy that misleads gene tree reconstruction
• B. Short intervals between successive speciation events leave little time for ancestral lineages to coalesce within each ancestral branch, so ancestral polymorphism persists through multiple speciations
• C. Rapid speciation always involves hybridization, creating chimeric gene trees that cannot represent either parental species
• D. Species that radiate rapidly have smaller effective population sizes, reducing the genetic variation needed for incongruence

Answer: (B)

ILS probability depends on the ratio of internodal time to effective population size, measured in coalescent units. In adaptive radiations, the intervals between successive speciation events are very short. There is little time for gene lineages in the ancestral populations to coalesce before the next split happens, so multiple ancient alleles get distributed across the new species and may sort into discordant gene trees. The same problem occurs with large ancestral population sizes, because more variation is present to sort discordantly.

Question 3

When a gene tree disagrees with the established species tree, one of the two trees is expected to be wrong — the goal of phylogenomics is to identify and discard the incorrect gene tree.

• T. True
• F. False

Answer: False

This is the central misconception the topic corrects. Gene trees and species trees are tracking different but real histories. A gene tree that disagrees with the species tree is not an error — it accurately reflects the genealogy of those gene copies, which happens to differ from the species branching pattern because of ILS, hybridization, or other processes. The goal of phylogenomics is not to discard discordant gene trees but to use the distribution of gene trees across the genome to estimate the species tree while explicitly accounting for the biological processes that cause discordance.

Question 4

The effective population size of an ancestral species is a key factor determining the probability of gene-tree incongruence at any given speciation node in the species tree.

• T. True
• F. False

Answer: True

Large ancestral populations maintain more genetic variation, giving gene lineages more distinct alleles that can be distributed discordantly among daughter species. The coalescent framework formalizes this: the expected time for two gene lineages to coalesce is proportional to 2Ne generations. A large Ne means coalescence takes longer, increasing the chance that lineages remain unsorted at the time of the next speciation event. This is why ILS is more severe in species with historically large populations, and why effective population size estimates are an important component of coalescent-based species tree inference.

Question 5

How does gene-tree incongruence from incomplete lineage sorting differ in its statistical signature from incongruence caused by hybridization, and why does this distinction matter?

Model answer: ILS produces random discordance: when a gene tree disagrees with the species tree, the two alternative discordant topologies occur at roughly equal frequencies because lineage sorting is a random process with no directional bias. Hybridization produces directional discordance: one discordant topology is enriched relative to the other because gene flow moves in a specific direction between specific lineages. This distinction matters because it allows researchers to diagnose the process causing discordance. If discordant topologies are roughly symmetric, ILS is the likely explanation. If one discordant topology consistently predominates, hybridization between specific lineages is implicated, which carries very different evolutionary implications.

The statistical signature of ILS versus hybridization is now used routinely in phylogenomics to test whether ancient gene flow occurred in a clade's history. This has produced major revisions in our understanding of, for example, human-Neanderthal-Denisovan relationships and the evolutionary history of many plant and animal groups.