Phylogenetics, Classification and Molecular Evolution

Animal phylogeny and tree of life

A phylogenetic tree is a hypothesis about how lineages are related, built from morphological and molecular characters. Relatedness is judged by the most recent common ancestor (MRCA) of two taxa, not by overall similarity, so two animals that look alike may lie far apart on the tree. sister group — the taxon that shares the most recent common ancestor with a given group. Example: if X and Y share a node that excludes Z, then X and Y are sister groups and Z is the outgroup.

The modern animal tree differs from older schemes built on body plan alone. Sponges (Porifera) and comb jellies (Ctenophora) are both non-bilaterian, so they are more closely related to each other than either is to bilaterians such as echinoderms. Cnidarians branch off before the bilaterian split, and within the Bilateria, protostomes and deuterostomes form separate clades. Among protostomes, Ecdysozoa (Arthropoda, Nematoda, and Tardigrada within Panarthropoda) are separated from Lophotrochozoa (Mollusca, Annelida). So Arthropoda is closer to Nematoda than to Mollusca.

Within vertebrates, a group that branches off before a clade is that clade's outgroup. Mammals are the outgroup of the reptile clade, and birds are nested inside dinosaurs, so birds are closer to dinosaurs than to crocodiles. Crocodiles are the sister group of dinosaurs and birds, which makes them closer to dinosaurs than to turtles.

Cetaceans are nested within the even-toed ungulates (Artiodactyla), with hippopotamuses as their closest living relatives. Among the usual comparison animals, pigs are the closest relatives, not elephants, seals, or zebras.

Taxonomic orders are identified by diagnostic traits. For example, Lepidoptera have scaled wings (moths), Hemiptera have piercing-sucking mouthparts (true bugs), Orthoptera have leathery forewings and enlarged hind legs (crickets), and Coleoptera have hardened forewings called elytra (beetles). Cone snails are gastropods, giant clams are bivalves, and octopuses and squids are cephalopods.

Some cell types are restricted to one phylum. Cnidocytes, which carry nematocysts for stinging, occur only in Cnidaria. Erythrocytes, myocytes, and neurocytes appear in several phyla. At the top level, Bacteria, Archaea, and Eukarya are the three domains. Archaea and Eukarya share a more recent common ancestor than either does with Bacteria, so Bacteria is the sister group to the Archaea-Eukarya pair in the standard tree.

Synonymous vs nonsynonymous substitutions

A substitution in a protein-coding sequence is synonymous (silent) if it does not change the encoded amino acid, and nonsynonymous (replacement) if it does. The genetic code is degenerate, with several codons for most amino acids, so many third-position changes are synonymous. degenerate code — the property that more than one codon specifies the same amino acid. Example: GCU and GCC both encode alanine, so a change from U to C at the third position leaves the protein unchanged.

Because synonymous changes do not alter protein sequence, two species can differ at a noticeable fraction of their DNA while their proteins are identical. Nonsense mutations, which create stop codons, and frameshift mutations, which shift the reading frame, both change the protein, so they cannot explain identical proteins with divergent DNA.

The ratio K_a/K_s (or d_N/d_S) compares nonsynonymous and synonymous substitution rates per site. A ratio below 1 indicates purifying selection, which removes amino acid changes. A ratio above 1 suggests positive selection. Under the neutral theory, synonymous changes accumulate at a rate close to the neutral mutation rate, and both classes of change accumulate roughly linearly with time, which is the expectation of a molecular clock.

Selection strength also varies across the parts of a gene. For a gene under purifying selection, exons are the most constrained and show the lowest divergence between sister species. Introns and flanking neutral regions evolve close to the neutral rate, so their divergences are approximately equal.

• purifying selection — removal of deleterious variants, which lowers nonsynonymous divergence. Example: a conserved enzyme-coding exon shows fewer changes than a neighbouring intron.
• neutral theory — most molecular change is selectively neutral, so substitution rates reflect mutation rates rather than selection.

Ancestral state reconstruction by parsimony

Ancestral state reconstruction estimates the character state at internal nodes of a tree. The parsimony principle chooses the state(s) that require the fewest evolutionary changes over the whole tree. Parsimony gives a minimum-change estimate and does not model branch lengths or rates. parsimony — the principle that the best explanation minimises the number of evolutionary changes. Example: if two sister tips share state A and a third tip has state B, the ancestral node of the first two is A, with one change on the lineage of the third tip.

When an ancestral state gives the same total number of changes under two alternatives, parsimony cannot prefer one. Both states are equally supported at that node.

For sequences, Fitch's algorithm is applied site by site. At each tip, the state set is the observed base. Moving up the tree, the node takes the intersection of its two child sets if they overlap, and otherwise the union, with one change counted. The root receives the set with the fewest changes. Applying this method to every site gives the ancestral sequence with the lowest total substitution count on the given topology.

Traits that look the same can have different histories. Homologous states are inherited from a common ancestor in which the state first arose. Analogous states arise independently in separate lineages. A pigmentation trait gained separately by two clades is analogous across those clades, even though the two species look alike. Shared absence inherited from a common ancestor is still homologous.

• homoplasy — similarity from independent origin or reversal, rather than common descent. Example: wings in bats and birds are analogous, not homologous.

Mitochondrial DNA markers

Mitochondrial DNA (mtDNA) in animals is small, usually maternally inherited, haploid, and lacks extensive recombination. It also evolves faster than most nuclear genes. These features make it a useful tool for phylogenetics, population genetics, and phylogeography.

The mitochondrial D-loop, or control region, is non-coding and weakly constrained by function. Mutations are tolerated and accumulate quickly, so it shows the highest sequence variation among the commonly compared mammalian regions. Hox genes, histone genes, and transcription factor binding domains are strongly conserved and vary little.

A species-identification marker must differ between species but be stable within them. The cytochrome c oxidase subunit I (COI) gene is the standard animal DNA barcode, and 16S rRNA is widely used to identify prokaryotic and some animal species. Microsatellites are highly polymorphic within species and are used for parentage, individual identity, and population structure, not for species identification. Immunoglobulins such as IgG are proteins and not genetic markers.

• DNA barcode — a short, standardised sequence used to assign an unknown specimen to a species. Example: COI sequences are compared with a reference library to identify an insect sample.

Historical colonization inferred from phylogeny

A colonization history is inferred by placing species from different regions onto a phylogeny. If a source region is basal, and a second region is nested inside a clade that includes the source, the second region was probably colonised through the intermediate area. Reversed or repeated placements reject that route.

For island or regional biota, the species set in each area can be mapped onto the tree. The lineages present show which branches reached each area and when. Summing the branch length spanned by a species set gives a measure of how much evolutionary history that area holds, which links colonization patterns to phylogenetic diversity.

• nested clade — a group that lies wholly within another group. Example: a Western Ghats clade nested inside an Eastern Ghats clade implies that the Western Ghats lineage descends from Eastern Ghats ancestors.

Zoonotic spillover tracing

A zoonotic spillover occurs when a pathogen jumps from a non-human reservoir to a new host, including humans. Testing a proposed route, such as a bat-to-pangolin transfer before infection of humans, requires evidence about the evolutionary relationships among viral genomes sampled from different hosts.

Comparative genomics identifies shared and divergent sequence features, including regions involved in host binding, across viruses from different hosts. Phylogenetic analysis places these viral sequences on a tree, showing which host's virus is the closest relative of the human-infecting virus and suggesting the direction of transmission.