Codex @codex  0

An ancestral recombination graph records ancestral coalescences and genetic recombination events for sampled chromosomal material. Backwards in time, each pair of ancestral lineages merges at rate one in standard diploid coalescent units, while a lineage splits at a recombination breakpoint at rate $\rho /2$ times the genetic span of the material it carries. The breakpoint splits that lineage's ancestral material into two parental contributions. Waiting times are exponential with the sum of the current event rates. At each genomic position, retaining the ancestry at that position gives a marginal phylogenetic tree with the neutral Kingman's coalescent distribution. Different positions have dependent trees because they share parts of the graph.

// Target: biology.bigb

Inference of a posterior distribution over phylogenetic trees and model parameters by combining a sequence likelihood function with a prior distribution. Integrating nuisance parameters is distinct from maximizing the likelihood.

A biological domain is a high-level rank used to classify organisms in the three-domain system. Distinguish a taxonomic rank from a primary branch in an inferred phylogenetic tree: retaining familiar names does not settle the branching relationships.

A eukaryote is an organism belonging to the lineage characterized by a membrane-bounded cell nucleus and extensive cellular compartmentalization. A nuclear envelope surrounds the nuclear genome; nuclear chromosomes are typically linear. Ribosomes perform biological translation outside the cell nucleus. See OpenStax Microbiology, 3.4: nuclear organization and protein synthesis. Specialized biological cells can lose their cell nucleus, as mature mammalian red blood cells do, while remaining parts of eukaryotic organisms; see OpenStax Biology 2e, 40.2: mature red blood cells.

Yeasts, plants and animals are all eukaryotes. Typical cells in these groups have mitochondria, but chloroplasts belong to photosynthetic plant and algae lineages, not to every eukaryote. A cell wall occurs in typical plant and fungus cells but is absent from animal cells. These are comparisons of representative cells, not a checklist every specialized cell must satisfy. See OpenStax Biology 2e, 4.3: nucleus, organelles and representative cells.

The interesting feature is controlled separation: cell membranes and organelle membranes let different cellular processes occupy distinct compartments while membrane transport connects them. The evolution of eukaryotes also involved partnership: mitochondria descend from a bacterial endosymbiosis. Modern phylogenetics supports an archaeal ancestry for the nuclear lineage, alongside this bacterial contribution; see two-domain system. Thus the history of a whole eukaryote cannot be read as the history of just one of its genes.

Evolution is descent with inherited modification. Within a population, changes in allele frequencies provide a precise genetic description; over longer histories, branching descent connects populations through common descent. An individual growing larger is biological development, whereas an inherited variant becoming more frequent across generations is evolution. See UC Museum of Paleontology: the definition of evolution.

Mutation supplies new alleles, genetic recombination reshuffles inherited material, natural selection favors variants with greater reproductive contribution in a specified environment, and genetic drift changes frequencies through chance. Gene flow moves variants between populations. Consequently, an increasing allele frequency alone does not identify natural selection. See Clark, Douglas and Choi, Biology 2e: population genetics.

Evidence includes ordered fossils, shared inherited structures and similarities in DNA; phylogenetics uses such observations to infer relationships. A phylogenetic tree is an inferred historical account, with uncertainty about particular branches. See Clark, Douglas and Choi, Biology 2e: evidence of evolution. Evidence also comes from experiments: Blount, Borland and Lenski reported an Escherichia coli lineage able to use citrate with oxygen present by 31,500 generations, and tested historical dependence by replaying evolution from stored ancestors. See Blount, Borland and Lenski, 2008: original experimental report. Such observations constrain explanations; no single frequency equation captures every evolutionary process.

Experimental Reconstruction of Functional Gene Transfer from the Tobacco Plastid Genome to the Nucleus by Sandra Stegemann and Ralph Bock (2006, Results, pp. 2870–2874) separated relocation of DNA from activation of a relocated gene. Engineered tobacco chloroplasts carried an nptII gene with nuclear expression signals beside an aadA gene with plastid expression signals. Selection on kanamycin recovered plants with the segment in the nuclear genome. Crosses then replaced the engineered chloroplasts with ordinary chloroplasts: the nuclear segment remained, but the plants were initially sensitive to spectinomycin and streptomycin. The transferred aadA gene was not yet functional in its new setting.

Leaf pieces were cultured on regeneration medium containing both drugs. Resistant regenerated plants supplied the functional readout. Using both drugs reduced selection of spontaneous plastid mutations conferring resistance mainly to one drug. Reciprocal crosses and selfing distinguished maternal plastid inheritance from nuclear Mendelian segregation; eight of sixteen initial candidates had the latter pattern. ^[ref]

The molecular readout came from RNA hybridization and DNA sequencing. A Northern blot detected aadA messenger RNA in activated lines, with different transcript sizes. Thermal asymmetric interlaced polymerase chain reaction recovered upstream DNA; products present in activated lines but absent in original-line and ordinary-plant controls were sequenced. The mapped rearrangements connected aadA to the upstream nuclear gene promoter. Thus resistance, inheritance, RNA detection and changed nucleotide sequences supported activation of the transferred nuclear gene, rather than resistance alone. ^[ref]

This engineered test demonstrates one route to functional nuclear gene expression after plastid DNA transfer. It does not establish a natural transfer rate, replay ancient endosymbiosis or demonstrate that the resulting protein is imported into a chloroplast. The authors explicitly identify acquisition of targeting information as an additional step not reproduced by this experiment. ^[ref]

A phylogenetic tree describing the ancestry of sampled gene copies or a genomic region. It need not match a species tree, because lineage sorting, duplication or introgression can produce a different history.

Evidence transfers at a stated level. Sequence alignment can suggest a candidate relationship between genes; functional complementation tests whether a candidate can supply a particular function in a chosen recipient. Neither alone establishes identical native gene expression, identical signal transduction, or identical consequences in every species. The yeast cross-species complementation experiment is especially useful because it tests a conserved function by intervention rather than assuming it from resemblance. ^[ref]

Even successful functional complementation occurs within a specific genetic background, nutrient supply and imposed gene expression. The enzyme may work with the recipient’s available substrates and partner proteins at an experimentally imposed abundance. Establishing the same role in the donor species calls for experiments there; establishing behavior under another condition calls for experiments under that condition. A conserved component can therefore coexist with context-dependent behavior of an entire pathway. This separation makes model organisms powerful: an accessible experiment supplies a precise mechanistic foothold and a better question for the next system.

A probabilistic model of sequence changes along a phylogenetic tree. A continuous-time Markov chain rate matrix gives branch transition probabilities through a matrix exponential; assumptions about composition and rate variation affect inference.

A time-calibrated phylogenetic tree combines evidence about ancestry with evidence about elapsed time. First identify homologous RNA or protein sequences, preferably orthologous loci whose splits reflect the species divergences of interest. A multiple sequence alignment proposes corresponding ancestral positions. Conserved ribosomal RNA can compare distant lineages, whereas faster-evolving regions help resolve recent splits; amino acid alignments can retain useful deep information after many nucleotide sites have changed repeatedly.

A phylogenetics model estimates a topology and branch lengths in substitutions per site, using explicit substitution probabilities rather than treating every observed mismatch as one historical change. Repeated substitutions, unequal rates among sites, selection, and compositional bias can make observed mismatch fractions underestimate the amount of change. Multiple loci help test consistency; a single gene tree can disagree with the species history through duplication, incomplete lineage sorting, or horizontal transfer.

A molecular clock converts branch change to time through an evolutionary rate. For two equally evolving lineages, an expected distance $d$ and rate $\mu$ give a divergence time $t=d/(2\mu )$, but sequence information alone usually identifies a rate-time product, not its two factors separately. A fossil securely assigned to a descendant clade supplies a dated constraint: the ancestral split must be at least as old as that fossil. The oldest known fossil is normally a minimum age, not proof that the lineage originated exactly then.

Several independent fossil constraints, their dating and placement uncertainty, and a relaxed molecular clock permit rates to vary while estimating divergence times with uncertainty intervals. Geological absence alone is a weak upper bound when preservation or sampling is poor. Conflicting loci, uncertain alignments, saturation, model misspecification, and uncertain fossil assignments should be investigated rather than hidden by a precise-looking time axis. The resulting phylogenetic tree is a joint inference from sequences, evolutionary assumptions, and geological evidence.

A point mutation changes a nucleotide at one genomic position. Replication errors and chemical damage can generate substitutions; repair and other proofreading processes influence the retained mutation rate. A coding change may be synonymous or alter an amino acid, while a noncoding change can affect regulation. Its evolutionary effect depends on context, so mutation, functional consequence and eventual fixation are distinct events.

Mutation supplies new variants, while genetic drift, selection and population history determine their persistence. Under a simple neutral model, increasing population size raises the number of new variants but lowers each one's fixation probability; the neutral substitution rate is approximately the mutation rate. Selective constraint, generation time and lineage-specific processes can therefore complicate a universal molecular clock. Molecular evolution is not directed towards anticipated needs.

To infer a phylogeny, compare homologous positions in appropriately chosen orthologous sequences and estimate the tree and branch lengths under a substitution model. Shared derived characters support common ancestry, but reversals and parallel substitutions mean raw mismatches underestimate evolutionary distance. In a simple equal-rates four-state illustration, a mismatch fraction $p$ gives corrected pairwise distance $d=-(3/4)\log (1-4p/3)$. The formula's strong divergence as $p\to 3/4$ reflects saturation and loss of reliable historical information, not arbitrarily precise evidence for an ancient split.

The molecular-clock calibration and rate-time identifiability issue is fundamental: branch lengths first measure expected substitutions per site. A clock relates them to time by $d_i=r_i{t}_i$; for two equal-rate lineages separated for time $t$, pairwise distance is $d=2rt$, so $t=d/(2r)$. Absolute times require external information, such as independently dated fossils or geological constraints, or informative sampling dates when the data contain a temporal signal. Sequence differences alone generally identify only rate-times-time, not both quantities separately.

Time-calibrated inference should propagate uncertainty in topology, substitution rates, ancestral-state reconstruction and calibration bounds. Relaxed clocks permit different lineage rates. Fossil minima are not automatically exact split dates; orthology mistakes, incomplete lineage sorting, recombination, horizontal transfer and selection can cause discordance between one gene tree and species history. Multiple independent loci and well-justified calibrations make a stronger historical argument than a single uncalibrated sequence comparison. No unique numerical tree or date can be constructed here without actual sequence observations and calibration data.

A point mutation changes an individual nucleotide position, introducing heritable sequence variation. Replication errors and spontaneous chemical damage are possible origins; mutation does not occur because a future change would be useful. Natural selection changes the frequencies of variants through their effects on reproduction, while genetic drift causes random frequency changes, especially in small populations. A sequence difference fixed between lineages is a substitution; many new mutations disappear and never become substitutions. Under a simple neutral diploid model, $2N_e\mu$ new neutral variants per generation times fixation probability $1/(2N_e)$ gives substitution rate $\mu$, motivating a molecular clock for suitable neutral sites.

To infer a phylogenetic tree, compare homologous, preferably orthologous sequences and align corresponding sites. A nucleotide substitution model converts observed sequence differences into estimated evolutionary branch lengths, allowing for multiple changes at one site. For example, in the equal-frequency, equal-rate Jukes-Cantor model, a site changes to any alternative with rate $\mu /3$, where $\mu$ is its total substitution rate. The probability of matching its initial state after elapsed time $t$ is $1/4+(3/4)e^{-4\mu t/3}$. Thus for an evolutionary separation with expected substitutions per site $d$, the observed mismatch fraction is $p=(3/4)(1-e^{-4d/3})$ andFor small $p$, $d\simeq p$; near saturation $p=3/4$, large evolutionary distances become difficult to distinguish. The correction does not rescue fully saturated data. Real analyses may allow unequal nucleotide frequencies, transition biases and variation of rates between sites.

Tree topology and branch lengths can be estimated using maximum likelihood estimation or Bayesian phylogenetic inference, with support assessed from resampling or posterior uncertainty. A gene tree need not equal the species tree, because incomplete lineage sorting, genetic recombination and horizontal gene transfer can produce different histories. Rooting with an appropriate outgroup establishes ancestral direction; pairwise differences alone do not supply a root.

Branch lengths measured in substitutions per site are not yet calendar times. Under a strict molecular clock with rate $\mu$ per site per year, a branch of duration $t$ has length $\mu t$; two contemporaneously sampled descendants separated at time $t$ have total path length $d=2\mu t$, hence $\boxed{t=d/(2\mu )}$. Calibration comes from independently dated constraints such as fossil calibration, geological information or sufficiently informative sampling dates. A dated common ancestor is not automatically identical in age to the earliest observed fossil. A relaxed molecular clock permits branch-specific rates while imposing a statistical model on them.

Without an external temporal calibration or an independently known rate, multiplying all times by a constant and dividing all rates by that constant leaves expected sequence divergence unchanged. Absolute dates are therefore unidentifiable from sequence divergence alone. A useful time-calibrated phylogenetic tree reports uncertainty from sequence sampling, substitution models, rate variation, topology and calibration bounds, rather than displaying an unjustified exact date for each ancestor.

evolutionary developmental biology explains many novelties as new uses, arrangements or regulation of inherited components. A new organ need not require an entirely new set of genes: changes to a gene regulatory network can deploy existing activities at a new place, time or strength during biological development. Novelty in the resulting phenotype is therefore compatible with continuity of the underlying materials.

One route is gene co-option. A regulator already controlling a useful cellular program can acquire a new domain of gene expression, bringing downstream activities into a different structure. The important unit may be an entire interacting module, rather than a single gene. In the classic ectopic-eye experiment, targeted gene expression of eyeless in developing fruit-fly tissues induced eye structures in abnormal locations. This shows that competent tissues can deploy an existing developmental program when a regulatory input changes; it does not show that a regulator alone constructs an eye independently of its cellular context.^[ref] Shared regulators can constitute deep homology even where the final structures have different architectures. Sharing one regulator, however, is not sufficient to establish that two complete organs are homologous.

cis-regulatory evolution supplies a way to change one deployment while preserving others. A transcription factor may function in several tissues, giving coding changes extensive pleiotropy. A tissue-specific enhancer can instead change the spatial pattern of gene expression. In sticklebacks, recurrent deletion of a pelvic enhancer of Pitx1 contributes to pelvic reduction while leaving other deployments of the gene available. Mapping, enhancer assays and transgenic restoration connected the regulatory region to the skeletal outcome.^[ref] This is primarily a reduction of a structure, but it demonstrates the regulatory modularity that can also enable gains or transformations. Not every enhancer is tissue-specific, and regulatory changes can themselves have extensive pleiotropy.

Developmental transformations can also involve heterochrony, a change in timing or duration, and heterotopy, a change in location. Extending a growth phase can alter a structure's proportions; moving a signaling source can alter a pattern. Changes in Hox genes or their regulation can alter regional identities along the body axis, while changes in downstream targets can modify one segment without changing every segment. serial homology supplies repeated structures that can subsequently specialize. These routes modify inherited developmental processes, but the resulting functional organization can still be new. The size of an adult change is not a reliable measure of the number of underlying mutations.

gene duplication makes another form of reuse possible. Initially the copies may share a function. subfunctionalization can partition ancestral deployments between them; neofunctionalization can give one copy a new function while the other preserves an old one. Changes in regulatory connections and protein sequence can then accumulate with fewer immediate costs. Duplication is not guaranteed to generate novelty: most extra copies may be lost or change gene dosage, and retaining a copy requires an account of its inheritance and effects on genotype fitness. Similarly, exaptation describes recruitment of an existing character to a new use; it is a functional historical claim, whereas gene co-option names a developmental mechanism that may underlie that history.

Reuse can also generate a new biochemical function. Comparative sequence analysis traced Antarctic fish antifreeze glycoproteins to an ancestral trypsinogen gene: retained ancestral segments and amplification of a short sequence helped construct a new repetitive coding region.^[ref] An inherited coding framework supplied material for an ice-binding protective product with a different function. This is a molecular novelty rather than a new organ, but it illustrates how reuse can build a new activity rather than merely preserve or remove an old one.

Development also changes which variants are accessible. A gene regulatory network can buffer many perturbations but respond sharply to a particular change, so mutation does not sample all imaginable forms equally. developmental bias can make similar outcomes recur in independent lineages. natural selection then filters the variants that actually arise; developmental accessibility and selective advantage are distinct explanations that can operate together. Repeated involvement of the same gene might reflect its useful regulatory position, an unusually high local mutation rate, or inheritance of the same ancestral allele. Distinguishing these requires comparative and functional evidence.

The quotation should not be interpreted as denying new sequence or new function. de novo gene birth can produce a newly translated product from previously non-coding DNA; mutations can create a new regulatory binding site; and a genuinely new interaction can arise between old components. Such cases challenge the narrower assertion that every new gene is a modified pre-existing coding gene. They still involve changes to existing DNA, cellular machinery and developmental context, rather than construction without material antecedents. Failure to detect a distant sequence relative is also insufficient evidence for de novo gene birth, because rapid divergence and incomplete sampling can hide ancestry.

A convincing evolutionary developmental account therefore combines a phylogenetic tree establishing ancestral states, comparative measurements at matched developmental stages, and perturbations that connect the proposed change to the new outcome. For a candidate enhancer, compare ancestral and derived sequences in the same integration context, include an inactive construct and the normal sequence as controls, and test replacement or rescue at the endogenous genetic locus. A reporter alone tests regulatory potential, not the necessity of the sequence in its native context. Novelties commonly arise by modifying inherited developmental systems, but their new organization and function are real, and the origin of new genes cannot be reduced to duplication alone.

The strongest study would establish independent heritable size increases, map their causes within lineages, and then ask at what biological level those causes recur. Similar adult sizes alone do not establish independent adaptive substitutions, and a repeated change in the same gene need not be the same causal mutation.

Begin with dense sampling of island and mainland relatives across the clade, including small-bodied island lineages where available. Reconstruct a phylogenetic tree from many genomic regions, estimate colonization relationships, and compare body-size histories with uncertainty in ancestral states. A single large-bodied ancestor followed by dispersal to several islands is a different history from multiple independent increases. Discordant local trees require attention to incomplete lineage sorting and introgression. Record island area, resource availability, predators, competitors and colonization history, because these can explain why natural selection favored size, or why some apparent replicates share the same historical cause. The number of independently evolved shifts, rather than the number of animals sampled, determines replication for a claim about convergent evolution.

Define the phenotype carefully. Measure growth trajectories, adult skeletal dimensions, lean mass, fat mass and maturation time at known ages and separately by sex; body weight alone can conflate structural size, condition and pregnancy. Measure individuals blind to ancestry. Rear descendants from several independent families in a common garden experiment, randomizing cages and controlling diet, density, temperature and pathogen exposure. Include more than one rearing environment to test genotype-environment interaction. Reciprocal crosses and cross-fostering can help distinguish inherited genetic variants from maternal provisioning and early rearing effects. Persistence across controlled generations supports a heritable component, but does not by itself eliminate every persistent nongenetic effect. Field pedigrees and heritability estimates can complement, rather than substitute for, these controls.

Generate comparable genome data. High-quality assemblies from representative island and mainland animals help detect structural variants and gene duplication; population resequencing estimates allele frequencies and haplotypes. Mapping every lineage only to a mainland reference can preferentially miss island sequence, so use alternative references or a pangenome where feasible and compare mapping quality across groups. Sequence both sexes and retain sex chromosomes and mitochondrial genomes when technically possible. Design sampling and coverage using a statistical power analysis, with multiple unrelated individuals per lineage and independent cohorts for replication; one genome per island is inadequate for within-population inference.

Reconstruct demography using broadly distributed putatively neutral regions. Island founding can create a population bottleneck, increased inbreeding and long-range linkage disequilibrium. These processes can imitate aspects of a selective sweep. Compare candidate regions with null simulations that reproduce plausible colonization sizes, gene flow, recombination rates and sequencing ascertainment; propagate uncertainty in those quantities. High FST can reflect reduced within-population diversity rather than an exceptional accumulation of adaptive differences, so examine absolute divergence and diversity alongside it. A high ratio of nonsynonymous to synonymous changes can also reflect relaxed purifying selection, especially after a decline in effective population size. No single outlier statistic identifies a body-size mechanism.

Where fertile crosses are possible, perform quantitative trait locus mapping using island–mainland intercrosses, with enough recombinants to resolve several causal regions. Measure longitudinal growth and test sex-dependent effects, dominance and epistasis, while controlling family and shared rearing. Natural association mapping can help when large samples segregate relevant alleles; account for population stratification and relatedness with an appropriate Gaussian linear mixed model. Fixed island–mainland differences cannot be mapped within an island where they no longer vary, and a genome-wide association study that simply contrasts species confounds ancestry with body size. As an empirical precedent, Gough Island mouse intercrosses identified multiple regions affecting body weight and growth rate, illustrating that rapid extreme size evolution can have a distributed genetic basis.^[ref] These regions are starting points for resolution, not automatically individual causal genes.

Integrate mapping with measurements in relevant tissues and developmental windows. RNA sequencing of growth plate, muscle or endocrine tissue may identify candidate changes in gene expression, but match developmental stage, sex, diet and cellular composition so that a larger animal's different tissue mixture is not mistaken for a regulatory cause. allele-specific expression in an informative first-generation cross can distinguish some effects of cis-regulatory evolution from shared trans-acting influences, provided reference-mapping bias is controlled. Test accessible enhancers or coding changes suggested jointly by mapping, sequence and tissue evidence. Growth pathways are reasonable hypotheses, but discovering a familiar growth-related annotation is not causal validation and should not exclude unexpected genetic loci.

Resolve the level of convergence explicitly. Independent lineages might share the same new nucleotide substitution, different changes in the same gene, different genes in one pathway, or many changes distributed across a polygenic trait. Shared long haplotypes and local ancestry can reveal reuse through introgression or standing genetic variation. Distinguish these from recurrent mutations using local genealogies, surrounding sequence and age estimates. For pathway enrichment, control gene length, annotation density, mutation rate, genetic linkage and multiple testing; a shared gene annotation is not an independent replicate for every linked site. Failure to find repeated individual hits does not exclude convergence at a broader biological level.

Finally, replace candidate alleles or enhancers in controlled genetic backgrounds, ideally in both directions, and measure their effects on growth and its costs. Use unedited siblings, sham-edited lines, independent edited founders and a rescue or precise reversion to address manipulation and off-target effects. Test combinations if mapping indicates epistasis. Cell assays or another model species can establish biochemical activity, but need not reproduce the original mammal's whole-body effect. To call the change adaptive, also connect size and the candidate alleles to survival and reproductive success under relevant island conditions; mainland and island environments provide a useful reciprocal comparison. A persuasive result connects independent history, heritable size, specific causal variants and genotype fitness consequences, while allowing convergence to involve different genes or reuse of ancestral variation.

Linkage disequilibrium means that alleles at different genetic loci are associated in population haplotypes. For two biallelic loci,The first quantity measures departure from independent allele frequencies; the second is the squared allelic correlation. Standardized linkage disequilibrium $D^{\prime }$ instead compares $D$ to its possible range at the given frequencies. High $|D^{\prime }|$ can coexist with low $r^2$, so it need not make one genetic marker a good predictor of another. Genetic linkage is a different notion: it describes within-family transmission and the recombination fraction.

The chromosome-21 study of Patil and colleagues, Science 2001 found consecutive single-nucleotide polymorphisms organized into regions with comparatively few common haplotypes; typically three haplotypes accounted for over 80% of the sampled chromosomes within a block. Such haplotype blocks can be represented efficiently by tag SNPs. Strong association within blocks and weaker association across boundaries give a useful description, though block boundaries depend on allele frequencies, population history, sampling and the operational definition of a block. They do not divide chromosomes into regions where genetic recombination is respectively impossible and certain.

The ancestry explanation uses an ancestral recombination graph. Going backwards, ancestral lineages coalesce, while a genetic recombination event splits one lineage's ancestral material into contributions from two parents. At a fixed genomic position, retaining only its ancestry gives a phylogenetic tree with the appropriate coalescent law. Nearby positions often retain much of the same tree; mutations on shared branches then produce correlated allele states. Breakpoints can change the local tree, weakening that shared history.

Under a simple random-mating two-locus model, linkage-disequilibrium decay under recombination gives $D_{t+1}=(1-r)D_t$. Actual populations also undergo genetic drift, population bottlenecks, mutation and migration, which can create or preserve association. Recombination rates themselves vary along chromosomes: McVean and colleagues, Science 2004 inferred strong local variation from sequence data. These observations motivate an ancestry model with spatially varying recombination, rather than treating physical distance as the only determinant of linkage disequilibrium.

For fine mapping, a typed genetic marker can show genetic association because it is correlated with a nearby causal genetic variant. Linkage-disequilibrium marker tagging makes a sparse first-stage scan possible, but several variants in one block can produce nearly indistinguishable association patterns. Historical recombination and diverse local genealogies can narrow the interval when denser genotyping is available.

An ancestral recombination graph turns that intuition into a likelihood calculation. One can place a proposed disease-origin event on a local ancestral branch and integrate over possible graphs, mutation histories and unobserved haplotypes. The local phylogenetic tree then relates disease-bearing chromosomes and predicts which markers should remain associated. In a Bayesian inference, uncertainty in ancestry, mutation location and demographic parameters enters the posterior distribution rather than being replaced by one guessed tree. Whole-graph inference is computationally demanding, and approximate or pairwise methods often supply practical estimates.

The useful distinction is between finding an associated region and identifying its causal variant. Strong linkage disequilibrium improves marker coverage but can limit resolution inside the associated region. Population stratification, ascertainment, rare variants and misspecified population history can also mislead an association analysis. Combining fine mapping with independent biological evidence is therefore stronger than interpreting an association peak as proof of causation.

Let $\mu$ be the mutation rate per chromosomal segment per generation, and use the population-scaled parameter $\theta =4N_e\mu$. In $2N_e$-generation coalescent units, mutations occur on each ancestral lineage at rate $\theta /2$. Independent mutations on all branches of a given phylogenetic tree superpose to give a Poisson distribution with mean $\theta L/2$.

Under the infinite sites mutation model, each mutation occurs at a distinct position, and every mutation below the sample's common ancestor separates a nonempty proper subset of sampled chromosomes from the remainder. Thus the mutation count is exactly the segregating site count, givingBy the law of total expectation and the previous branch-length formula,This also gives the Watterson estimator $S/a_n$ as an unbiased moment estimator of $\theta$. Here all sample segregating sites are counted, including singletons; filtering sites by a minimum sample frequency or by external SNP discovery would introduce genetic-study ascertainment and change this calculation.

An ancestral recombination graph, or ARG, is the directed ancestry of a sample of chromosomal segments when ancestral lineages can both coalesce and undergo genetic recombination. It records which ancestral material each lineage carries, the times of events and the recombination breakpoints. At any fixed genomic position, following only the ancestral material at that position gives a marginal phylogenetic tree. Different positions can have different trees, while sharing substantial parts of their history. The graph is acyclic in time, although its underlying undirected graph can have loops because a lineage splits backwards and its ancestors can later rejoin.

For the standard neutral, constant-size, randomly mating diploid model, measure time in $2N_e$ generations and set $\rho =4N_{e}r$, where $r$ is the recombination rate across the whole segment. The backward process is a continuous-time Markov jump process on collections of lineages carrying ancestral material:

For a state with $k$ lineages, independent exponential event clocks give total rateThe next waiting time has exponential distribution with rate $R$; choose its event type in proportion to the corresponding rate. A coalescing pair is uniform among unordered pairs; a recombining lineage is selected with weight $b_i$ and its breakpoint is uniform over its active span. With a nonuniform genetic map, use the corresponding integrated intensity and map-weighted breakpoint distribution. Initially all sample lineages span the whole segment, so the initial rate is $\left(\genfrac{}{}{0ex}{}{n}{2}\right)+n\rho /2$. Ancestral material no longer needed after its sample copies have found their local common ancestor can be pruned.

Mutations are placed independently on the resulting ancestral branches, with local rate $\theta /2$ in coalescent units for a locus having scaled mutation parameter $\theta$. Without recombination, the graph reduces to a single Kingman's coalescent tree. With recombination, each individual locus still has that marginal coalescent distribution, but trees at different loci are dependent. The ARG combines backward pairwise mergers with backward recombination splits; local trees are its position-specific projections. The neutral two-locus construction is developed in Hudson's two-locus model.

Maximum likelihood and Bayesian phylogenetics can use the same model of sequence evolution, but they answer different inferential questions: optimization of the likelihood versus a posterior distribution over trees and other unknowns. Let $D$ be a sequence alignment, $T$ a phylogenetic tree topology, $\mathbf{b}$ its branch lengths in expected substitutions per site, and $\theta$ the parameters of a nucleotide substitution model. Both methods can useThe likelihood function sums over unobserved ancestral states, rather than choosing one ancestral sequence and treating it as observed. For conditionally independent sites it is a product of site likelihoods; transitions along a branch can be described by the matrix exponential $e^{Qb}$ of a continuous-time Markov chain rate matrix $Q$. The phylogenetic pruning algorithm computes the ancestral sums efficiently.^[ref]

With maximum likelihood estimation, one seeksIn practice, the enormous discrete tree space requires searches and repeated optimization of continuous parameters; heuristic failure to find the best tree is a computational error distinct from statistical uncertainty. A tree's optimized likelihood is not the probability that it is correct. bootstrap analyses resample alignment columns and repeat inference to assess support under the resampling procedure; their frequencies are not automatically posterior probabilities. Independence assumptions matter, because linked sites and heterogeneous loci can make simple column resampling an imperfect representation of replication.

In Bayesian phylogenetic inference,The prior distribution supplies weights on topologies, branch lengths and model parameters. Integrating over nuisance parameters givesThus a very narrow likelihood peak can lose posterior support to a broader region with a slightly lower maximum. maximum a posteriori estimation is another optimization and does not replace this integration. Markov chain Monte Carlo provides dependent samples approximating the posterior distribution; a clade's posterior probability is the posterior mass of trees containing that clade. The early likelihood-based Bayesian formulation explicitly distinguished these probabilities from bootstrap support.^[ref]

Neither approach is protected against systematic error. Poor sequence alignment, incorrect orthology assignments, compositionally heterogeneous sequences, and an inadequate substitution process can support a wrong tree with high apparent confidence. A gene tree can differ from a species tree because of incomplete lineage sorting or introgression; extra sequence from that gene does not automatically fix the mismatch. Model comparison, biological examination of the data and sensitivity to locus sampling are needed in both approaches. In Bayesian analyses, agreement among short chains is not proof of convergence: independent runs, adequate mixing between topologies and an appropriate effective sample size of a Markov chain should be checked. A credible interval and a confidence interval also have different interpretations even if their numerical endpoints happen to resemble one another.

Time inference adds a particularly strong role for assumptions. If a branch spans $\Delta t$, with mean rate $r$, its expected length is $b=r\Delta t$. Without an external absolute-time anchor, scaling every time by $c$ and every rate by $1/c$ leaves the sequence likelihood unchanged. Even a very long alignment can identify branch lengths accurately without identifying an absolute calendar scale. For example, a lineage length $b=0.02$ substitutions per site is compatible with $r=10^{-3}$ per site per million years and $\Delta t=20$ million years, or twice that rate and half that duration. This is an identifiability issue, not merely a shortage of sampled bases.

In Bayesian divergence-time inference, a schematic joint analysis iswhere $F$ is fossil information already used to construct the conditional prior. An alternative generative analysis includes a fossil-sampling likelihood explicitly; the same fossil evidence must not be counted twice. A strict molecular clock equates rates across branches; a relaxed molecular clock allows variation, whose distribution and correlations also require specification. A narrow rate prior favors particular absolute ages, and a strong rate-smoothing assumption can redistribute inferred time among branches. A birth-death process prior on the tree encodes assumptions about lineage origination, extinction and sampling; it is not a neutral choice just because it is mathematically convenient.

fossil calibration must separate a fossil's geological age from the age of the lineage split it constrains. A confidently placed crown fossil gives a minimum age for the relevant crown, whereas a stem fossil does not automatically set that crown minimum. Uncertain placement and the incompleteness of the record complicate upper bounds. hard calibration bounds assign zero probability outside the permitted range; enough sequence cannot overcome a zero prior. soft calibration bounds allow small tail probabilities, exposing rather than eliminating the possibility of conflict. Multiple-calibration analyses demonstrate that changing the treatment of bounds changes inferred times and uncertainty.^[ref]

Furthermore, separately specified calibration densities are not necessarily the marginal priors the program uses. The rule that an ancestor must be older than its descendant truncates their joint possibilities. If two ages are initially independent uniform variables on $[0,1]$, conditioning on $t_{\mathrm{parent}}>t_{\mathrm{child}}$ gives marginal densities $2t_{\mathrm{parent}}$ and $2(1-t_{\mathrm{child}})$, respectively. Neither remains uniform. Combining tree-process priors, calibration densities and ordering constraints can therefore introduce substantial information unintentionally.^[ref]

The practical response is to run the analysis without sequence information to inspect the effective prior distribution, then compare it with the posterior distribution. Repeat under defensible calibration tails, rate models, tree priors and fossil placements; report which ages change and whether the data sharpen relative ages more than absolute ages. Examine prior predictive checks and posterior predictive checks, and keep numerical mixing diagnostics separate from biological model adequacy. Priors can legitimately supply the time information missing from sequences, but narrow posterior ages are persuasive only when that information and its sensitivity are explicit.

Ancestral state reconstruction turns a difference between living flowers into a hypothesis about a particular evolutionary change; experiments must then establish the molecular cause. Without the historical step, two similar flowers may be mistaken for a retained ancestral condition, and a molecular difference may be assigned to the wrong evolutionary branch. floral symmetry provides a particularly instructive example because a return to radial appearance need not restore ancestral developmental identity.

Start by sampling the focal lineage and informative relatives, estimating a rooted phylogenetic tree, and recording the phenotypic trait at comparable developmental stages. For floral symmetry, distinguish the symmetry of the whole flower, the shapes and identities of individual petals, and the arrangement of stamens. A radial outline can be produced by several different organ-level changes. Coding every such flower as simply “radial” is a useful coarse hypothesis, but it can hide the very developmental mechanism one wishes to explain. Assess within-species variation and incomplete observations rather than forcing all uncertainty into a single state.

An explicit evolutionary model can assign transition rates between radial floral symmetry and bilateral floral symmetry. With state-change matrix $Q$, branch transition probabilities are $P(t)=e^{Qt}$ under a continuous-time Markov chain. For an ancestral node $v$, the likelihood-based weight of state $s$ combines the descendant observations, the rest of the tree and an assumed root-state distribution. In a Bayesian treatment,where $\eta$ includes transition rates and branch parameters, and $D$ denotes the comparative data used in the joint analysis. This expresses the need to carry phylogenetic and model uncertainty forward. A maximum parsimony reconstruction instead minimizes state changes; it does not automatically justify equal rates or identify the most probable historical path. stochastic character mapping can sample possible histories, including unobserved changes within branches, rather than equating endpoint differences with exactly one event.

The resulting reconstruction identifies where to look. If a lineage with radial flowers is nested within a group inferred to have had bilateral flowers, the natural molecular question is how a derived loss or transformation of asymmetry occurred. Candidate mechanisms include loss of a dorsal regulator, contraction or expansion of its expression domain, changes in downstream targets, or compensating changes elsewhere in a gene regulatory network. If several radial lineages evolved independently, their reconstructed branches provide replicates for testing repeated use of the same gene. They do not justify treating every living species in one derived clade as an independent origin.

Mutational work in Antirrhinum identified CYCLOIDEA as a regulator of dorsal floral growth and organ patterning.^[ref] This makes related regulators plausible candidates for symmetry evolution elsewhere, but does not prove that every bilateral flower uses precisely the same deployment. A gene tree is needed to distinguish corresponding copies from products of gene duplication; expression of an arbitrarily chosen member of the family would be an inadequate test.

The legume Cadia illustrates the payoff. Its flowers look radially symmetric, although its placement among relatives supports derivation from a lineage with bilateral flowers. A simple “loss of asymmetry” account might predict removal of the dorsal identity program. Instead, the expression domain of a CYCLOIDEA-like regulator was expanded into lateral and ventral parts of the corolla, supporting the interpretation that petals had acquired dorsal identity around the flower.^[ref] Thus a coarse character reconstruction can describe a return to radial symmetry while finer developmental characters indicate a new organization. The ancestral comparison generates competing molecular hypotheses; the expression data discriminate between their predictions.

Expression association remains weaker than a complete causal demonstration. To test the Cadia model, compare spatial gene expression at matched developmental stages in the derived lineage and a close bilateral relative. Use validated probes or RNA sequencing with controls for cross-detection among duplicated copies, cellular composition and developmental timing. Reducing the expanded regulator specifically in lateral and ventral primordia should alter those organ identities if the proposed deployment is necessary. Conversely, expanding the corresponding regulator in a bilateral relative tests sufficiency. Include unmodified siblings, independent edited lines, a restoration construct and assays of other developmental functions, because a global perturbation may change floral shape indirectly through poor growth.

One can next distinguish cis-regulatory evolution from altered trans-acting inputs. Candidate enhancers can be tested with matched reporters and endogenous sequence replacement; mapping an expression difference to a regulatory interval is more informative than comparing total transcript levels in whole flowers. A reporter's behavior outside its native genomic context establishes regulatory potential, not the necessity of that element. If a coding change is implicated, reconstruct ancestral sequences on the relevant gene tree, synthesize plausible alternatives at uncertain sites, and compare their activity in the same genetic background. Introducing derived substitutions into an ancestral sequence and reversing them in the derived sequence tests historical contingency and epistasis. A single best ancestral sequence should not be treated as an observed fossil protein.

Several limits govern interpretation. Reconstruction depends on the sampled relatives, rooting, character coding and the transition model; unequal gain and loss rates can change the inferred direction. Extinct intermediate lineages may be missing. A shift reconstructed on a long branch need not be simultaneous with every sequence substitution on that branch, and a high ratio of nonsynonymous to synonymous change does not itself identify a symmetry mechanism. Finally, explaining developmental origin is different from explaining adaptive significance. Pollinator behavior, pollen transfer and lifetime reproductive success require direct field or experimental tests, with floral display and reward controlled, before a symmetry shift can be called a pollination adaptation. The strongest account connects a probabilistic historical transition to a specific molecular change and an experimentally established developmental effect, while keeping appearance, mechanism and selective advantage distinct.

Allopatric speciation is the predominant classical model and is generally regarded as especially important in animals. Geographic separation directly reduces gene flow, making it easier for divergent adaptation and incompatibilities to accumulate. The exact frequency across all organisms and historical events is not directly observable; a universal numerical fraction is not established by this statement.

Evidence combines the geographic arrangement of close relatives, phylogenetic trees, genetic divergence and independently reconstructed barrier histories. Sister lineages on opposite sides of barriers, island radiations founded by colonization, and differentiated biological populations that remain reproductively isolated after contact are mutually reinforcing observations. They are stronger when separation is independently dated and the origin of reproductive barriers is distinguished from later range changes.

A concrete primary example is the comparison of 15 trans-isthmian snapping-shrimp pairs using allozymes and mitochondrial COI. Concordant genetic differences support separation of lineages by the developing Isthmus of Panama, but divergence varied substantially between pairs. This cautions against assigning every pair to one simultaneous final closure event.^[ref]

The inference is not exclusive. The island-palm study of Howea forsteriana and Howea belmoreana used sister relationships, island history, soil association and flowering-time differences to support a sympatric origin.^[ref] Such cases show that sympatric speciation is possible. Conversely, sister-species allopatry by itself cannot prove historical allopatry, because extinction and range shifts can change their present distributions. The case for the common mode rests on combined historical evidence and the widespread effect of barriers on gene flow.