MLB7052.LT: Modern Methods in Molecular Ecology
TalTech / UT Institute of Technology
Loeng 10 — neutraalne ja adaptiivne molekulaarne muutlikkus (neutraalne ja adaptiivne molekulaarne muutlikkus).
Every marker you have used so far was assumed to be neutral. This lecture asks what that assumption actually buys you, when it fails, and what you do about the variation that is not neutral.
Key terms are given with their Estonian translation. Eestikeelne materjal: A. Heinaru, Geneetika, evolutsioonigeneetika peatükk.
The genetic rescue toolkit: bioinformatics, molecular biology, advanced reproduction, engineering — from genomic insight through to de-extinction.
…and we made maps. Even a simple one is informative (ka lihtsad kaardid võivad olla informatiivsed).
In the comparative biogeographic approach, relationships are reconstructed across multiple clades of co-distributed species, and the patterns they share are attributed to shared events.
In some frameworks time is treated explicitly; in others only the branching order matters.
Meta-analysis then asks whether a factor — pelagic larval duration, fecundity, population size — predicts an effect size such as a difference in .
| 1 | Forces of evolution and the fate of new mutations |
| 2 | The neutral theory and the molecular clock |
| 3 | Detecting selection in DNA sequences |
| 4 | Which molecular markers are neutral? |
| 5 | Adaptive molecular variation in nature |
| 6 | Quantitative traits, gene expression and conservation |
Synthetic theory of evolution (sünteetiline evolutsiooniteooria): Fisher, Wright, Haldane, 1920s–1930s.
| Force | Eesti keeles | What it does |
|---|---|---|
| Mutation | mutatsioon | The only source of new alleles. Spontaneous rate ~– per nucleotide per generation |
| Natural selection | looduslik valik | Non-random differences in survival and reproduction among genotypes |
| Genetic drift | geenitriiv | Random change in allele frequency from the sampling of gametes; strongest in small populations |
| Gene flow | migratsioon, geenivool | Moves alleles between populations; homogenises them and spreads new variants |
Hardy–Weinberg equilibrium (Hardy-Weinbergi tasakaal):
In a large, randomly mating population with no mutation, selection, drift or migration, allele and genotype frequencies stay constant. Departures — or change over time — point to one of the four forces.
Red fox fur colour, Canada 1830–1930: red , reddish-black , silver . Trappers preferentially took the valuable silver pelts, and the frequency of the allele declined.
Mutation rates vary almost 1000-fold between species — from per nucleotide per generation in some unicellular organisms to ~ in primates.
Only 1–2% of a large eukaryotic genome encodes structural proteins, so most new mutations land in the vast non-functional regions.
A new mutation may be:
Fitness (kohasus) is the ability to survive and reproduce; a common measure is lifetime offspring number.
The absolute fitness of a genotype is the proportional change in its abundance over one generation attributable to selection:
Relative fitness , and the selection coefficient .
Fitness components trade off: more offspring often means lower survival.
The difference between the fitness of the most fit genotype and that of the average genotype in the population.
| Low genetic load | Reproductive fitness is similar among all organisms; most are likely to survive and reproduce |
| High genetic load | Greater variance in fitness: few carry the most-fit genotype, the majority have much lower reproductive fitness |
Fisher’s (1930) treatment predicted that any genetic load would decline, reducing population genetic variability to low levels — and Drosophila studies showed that many mutations are indeed deleterious.
In the 1920s, mathematical treatments of natural selection — culminating in Fisher’s The Genetical Theory of Natural Selection (1930) — reinvigorated evolutionary biology. This became The Modern Synthesis (sünteetiline evolutsiooniteooria), a pervasive selectionist view.
Its expectation was clear: genetic variability in most species would be low, with most individuals homozygous for the same allele at most loci, because mutation-produced variation would create a heavy genetic load (Muller 1950).
| 1859 | 1900 | 1920–30s | 1950 | 1966 | 1968 | 1973 |
| Darwin | Mendel rediscovered | Fisher, Wright, Haldane | Muller — genetic load | Lewontin & Hubby; Harris | Kimura | Ohta |
Gel electrophoresis of allozymes (Lewontin & Hubby 1966; Harris 1966) revealed far more variation than anyone expected:
How is so much variation maintained?
| Selectionist answer | Balancing selection (tasakaalustav selektsioon) actively keeps alleles in the population |
| Neutralist answer | Most variants are selectively neutral (selektiivselt neutraalne) and drift randomly towards fixation or loss |
Each generation is a finite random sample of the parents’ gametes, so allele frequencies wander by chance — random genetic drift (juhuslik geenitriiv).
20 simulated populations starting at . After 20 generations is fixed in 7 and lost in 5.
Tympanuchus cupido in Illinois (Westemeier et al. 1998, Science).
| Mode | Eesti keeles | Example |
|---|---|---|
| Directional | suunav valik | Galápagos finches: mean beak depth rose within two years of the 1977 drought |
| Stabilizing | stabiliseeriv valik | Gall flies: wasps kill larvae in small galls, birds in large ones |
| Disruptive | lõhestav valik | African seedcrackers: small and large mandibles each handle a seed type |
| Balancing | tasakaalustav selektsioon | Maintains polymorphism — heterozygote advantage, frequency dependence, seasonal change |
| Purifying | puhastav selektsioon | Removes deleterious mutations — the commonest selection at the molecular level |
Motoo Kimura, 1968.
Most evolutionary changes at the molecular level, and most variation within and between species, are due to random genetic drift of selectively neutral mutant alleles.
Motoo Kimura (1924–1994)
The rate of neutral substitution is independent of population size — cancels out. If is constant, substitutions accumulate steadily: a molecular clock (molekulaarne kell).
Rates differ between genes and regions because the fraction of mutations that is neutral differs — functional constraints.
Amino-acid differences accumulate linearly with divergence time.
Kimura’s neutral theory unifies:
Intrapopulation polymorphism is just a random walk of variants on their way to fixation or loss.
Polymorphism and divergence are therefore two phases of the same process — which is exactly what the HKA and McDonald–Kreitman tests exploit.
The more important a protein or region, the slower it evolves — functional constraints (funktsionaalsed piirangud).
Constraint is not noise. It is the property that makes a marker suitable for a particular taxonomic depth.
Substitutions per site per years in different parts of a gene:
| 5.1 | α-globin pseudogene ψα1 (pseudogeen) |
| 2.6 | 3rd codon positions — mostly synonymous (sünonüümne asendus) |
| 0.7 | 1st + 2nd codon positions — mostly non-synonymous (mittesünonüümne asendus) |
Fastest where mutations matter least; slowest at non-degenerate sites — exactly as the neutral theory predicts.
Sequencing 11 Adh alleles (Kreitman 1983) revealed 43 polymorphic sites.
Only one of them — Lys ↔︎ Thr at codon 192 — changes the protein: the classic Fast/Slow allozyme. The rest lie in introns, UTRs, flanking regions, or in synonymous codons.
Allozymes reveal only a small fraction of the DNA variation that is actually there.
Point mutations in human β-globin (146 aa): variants are known at 9 amino-acid positions, yet only HbS, HbC and HbE cause disease.
Many replacements are effectively neutral — detectable in the protein or the DNA, invisible in the phenotype.
HbS (Glu6Val) is the exception that proves the rule: homozygotes have sickle-cell anaemia, but heterozygotes resist malaria — overdominance maintained by balancing selection.
(A) Kimura’s neutral theory, 1960s. (B) Ohta’s nearly neutral theory, 1970s, with selection coefficients graded from strongly deleterious through neutral to advantageous.
A mutation behaves as neutral when — drift outweighs selection.
Slightly deleterious mutations can therefore drift to fixation in small populations but are purged in large ones, which makes the substitution rate depend on after all.
| Model | Explains evolution by |
|---|---|
| Neutralist | mutation + random genetic drift + purifying selection |
| Selectionist | advantageous (positive) selection + balancing selection |
| Nearly neutral | borderline mutations, whose effects lie between the selected and the neutral classes |
These are not three camps to pick from. They are three regions of the same distribution of selection coefficients, and which one dominates depends on the gene and on .
The MK test is the standard statistical test of neutrality versus selection using DNA sequences. It needs sequences of several alleles from within a species and at least one sequence of the same gene from a different species.
For a protein-coding locus, the neutral theory predicts that the ratio of non-synonymous to synonymous changes within a species equals the same ratio between species.
| Fixed between species | Polymorphic within species | |
|---|---|---|
| Synonymous | ||
| Non-synonymous |
| Negative selection — disadvantageous mutations inflating polymorphism (negatiivne selektsioon) | |
| The neutral expectation (neutraalsus) | |
| Positive selection — adaptive amino-acid fixations (positiivne selektsioon) |
The neutrality index quantifies the departure:
→ negative selection; → positive selection; → neutral. (Smith & Eyre-Walker 2002) estimates the fraction of amino-acid substitutions fixed by positive selection.
D. melanogaster, D. simulans, D. yakuba:
| Fixed | Polymorphic | |
|---|---|---|
| Non-synonymous | 7 | 2 |
| Synonymous | 17 | 42 |
vs
, , Fisher’s exact test .
An excess of fixed amino-acid differences between species → adaptive protein evolution at Adh.
Tähtis
Caveat: slightly deleterious polymorphisms inflate and make an underestimate. This is why rare variants are often excluded before the test is run.
| Signature | What it measures |
|---|---|
| dN/dS () | Non-synonymous to synonymous rate ratio along a phylogeny: purifying, neutral, positive |
| Tajima’s | Site-frequency spectrum: excess of rare variants (sweep or expansion); excess of intermediate variants (balancing selection or bottleneck) |
| outlier scans | Loci far more (or less) differentiated between populations than the genome-wide background → local adaptation |
| Haplotype tests | Long unbroken haplotypes at high frequency (EHH, iHS) reveal recent sweeps |
Tähtis
Demography — bottlenecks, expansions, population structure — can mimic every one of these. Always compare against a genome-wide neutral background.
| Neutral markers | Until recently molecular ecology focused on polymorphic markers not directly affected by selection. A large body of evidence testifies to the safety of that assumption for most markers in most situations. |
| Adaptive variation | Traits of functional significance, likely to be under some form of selection: behaviour, morphology, life history — quantitative characters underpinned by alleles at many genes. |
Assumptions of marker neutrality are central to much of molecular ecology. They may sometimes be violated, and in any case they omit the adaptive variation in natural populations.
So: are the markers really always neutral?
Allozymes (allosüümid) — forms of one enzyme encoded by different alleles of a gene, separated by gel electrophoresis.
Environmental selection for optimum enzyme performance acts on allozyme variants, particularly in ectotherms.
Fundulus heteroclitus shows a latitudinal cline of lactate dehydrogenase (LdhB) alleles that tracks ocean temperature along eastern North America.
frequency falls with latitude along eastern Australia.
Between 1979–82 and 2002–04 the whole cline shifted about 4° of latitude polewards as the climate warmed (Umina et al. 2005, Science).
A “neutral” allozyme responding to temperature over twenty years is selection, not drift.
This is the same gene as the silent-polymorphism and MK examples — one locus, three lessons.
mtDNA became popular for its technical ease, supposed clonality, near-neutrality and clock-like substitution rate. However:
Märkus
Evidence keeps accumulating that strict neutrality of mtDNA is invalid. That does not make mtDNA useless; it makes “mtDNA is neutral” a claim you have to defend for your system rather than assume.
Spiroplasma protects Drosophila neotestacea against a sterilising nematode (Jaenike et al. 2010, Science).
mtDNA diversity can then reflect symbiont dynamics, not neutral demography.
Factors contributing to fitness constitute only part of the adaptive variation in natural populations.
Much morphological variation may have little immediate relation to fitness — but could become important if environmental circumstances change.
Neutral molecular markers are generally poor indicators of adaptive variation. That is the whole reason this section exists.
Dark mice on black lava, light mice on pale rock (Chaetodipus intermedius). Coat colour matches substrate under strong selection by visual predators.
Pogonus chalceus has short-winged and long-winged populations — wing length directly controls dispersal capacity.
Found exclusively in salt marshes, largely restricted to the coastal tidal zones of the Atlantic and Mediterranean.
Mean population wing size is tightly correlated with the frequencies of allozymes of the nuclear-encoded mitochondrial NADP⁺-dependent isocitrate dehydrogenase (mtIdh).
The two reflect similar selection pressures from different hydrological regimes: regular short tidal inundation versus unpredictable seasonal inundation.
Short-winged-associated mtIdh-D haplotypes cluster within the long-winged mtIdh-B haplotypes.
The same genealogy supports two opposite stories. Which one you believe depends on which summary statistic you privilege.
A key adaptive physiochemical trait reconstructed in an extinct species (Campbell et al. 2010).
Haemoglobin binds and carries O₂, but offloading it to respiring cells is hampered at low temperature, because haem deoxygenation is endothermic — Hb–O₂ affinity rises as temperature falls.
Mammoth haemoglobin, “resurrected” from ancient DNA sequence and expressed in the lab, carries amino-acid substitutions that make O₂ binding less temperature-sensitive — an adaptation to cold.
A resistance allele arises once by mutation. Gene flow — or accidental transport — carries it to new populations, where insecticide selection drives it to high frequency.
In Culex pipiens, amplified esterase resistance alleles spread worldwide from a single origin (Raymond et al. 1991).
Neutral markers measure the gene flow; adaptive alleles ride on it.
Ram Mountain, Alberta (Coltman et al. 2003, Nature).
Harvest selects like a predator — but on the largest individuals, which no natural predator does.
Where G. fortis and G. fuliginosa co-occur, their beaks differ strongly; on islands where each lives alone, they are similar. Competition drives divergent selection on beak size.
During the 2004–05 drought, an HMGA2 haplotype for small beaks rose in G. fortis as the larger G. magnirostris took the big seeds (Lamichhaney et al. 2016).
Ecological interactions leave signatures at specific loci.
Most traits that vary among individuals result from the combined effects of different alleles at multiple genes — particularly traits that bear upon fitness. In the mouse, at least 650 genes have some phenotypic effect on growth.
Quantitative trait analysis usually requires laboratory breeding, but in wild populations relatedness can be estimated with molecular markers instead.
| Term | Eesti keeles | Meaning |
|---|---|---|
| fenotüübiline varieeruvus | Total phenotypic variance, genetic and environmental | |
| aditiivne geneetiline varieeruvus | Additive genetic variance — loci whose allele effects add up | |
| päritavus | Heritability, | |
| pleiotroopsus, aheldustasakaalutus | Additive genetic covariation of several traits | |
| G-maatriks | Matrix of and — the directions adaptation can take | |
| dominantsus | Non-additive variance, mostly dominance |
| response — change in the mean of the next generation | |
| heritability () | |
| selection differential — mean of selected parents minus population mean |
Selection changes the next generation only through the heritable part of the variation. A strong selection differential on a trait with produces nothing.
Galápagos finches: offspring of drought survivors inherited deeper beaks — evolution observed within one generation (Grant & Grant).
Weber, Peterson & Hoekstra 2013, Nature.
A complex behaviour splits into a few genetic modules.
Gene expression has been hypothesised to be of adaptive importance — and heritable variation that affects fitness is necessary for evolution by natural selection.
Adaptive expression differences have been found in single-gene studies; microarrays and now RNA-seq assay many loci at once and can test the hypothesis rigorously.
Much variation in expression under a given condition has a genetic basis — shown in yeast, Drosophila, mice and humans.
Neutral expectation: expression divergence accumulates with time like neutral sequence divergence. Deviations suggest selection.
Four scenarios of genomic response to an environmental challenge across taxa — each supporting a different evolutionary inference.
| Timescale | What is happening |
|---|---|
| Minutes–weeks | Stress, then acclimation (aklimatiseerumine) — phenotypic plasticity (fenotüübiline plastilisus), often via gene expression |
| Generations | Adaptation (kohastumus) — change in allele frequencies |
To tell them apart: common-garden and reciprocal-transplant experiments, and eQTL mapping.
The two look identical in a single field sample. They are distinguished by experimental design, not by sequencing depth.
Small populations enter a feedback loop: inbreeding depression (inbriidingdepressioon), drift and loss of genetic diversity → lower fitness → even smaller .
Management units should integrate both.
| 1 | Four forces — mutation, selection, drift, gene flow — shape molecular variation |
| 2 | Neutral theory: most molecular change is drift of neutral mutations; neutral substitution rate = mutation rate → molecular clock |
| 3 | Functional constraints explain why rates differ among genes, sites and protein regions |
| 4 | Polymorphism vs divergence (MK test, , ) and other signatures detect selection — mind demography |
| 5 | Popular “neutral” markers (allozymes, mtDNA) are not always neutral |
| 6 | Adaptive variation ranges from single genes (Mc1r) to polygenic traits (, QTL) and plastic gene expression |
| English | Eesti keeles |
|---|---|
| Mutation | mutatsioon |
| Natural selection | looduslik valik |
| Genetic drift | geenitriiv |
| Gene flow | geenivool |
| Fitness | kohasus |
| Genetic load | geneetiline koormus |
| Fixation | fikseerumine |
| Effective population size | efektiivne populatsiooni suurus |
| Modern Synthesis | sünteetiline evolutsiooniteooria |
| English | Eesti keeles |
|---|---|
| Neutral theory | neutraalsusteooria |
| Selectively neutral | selektiivselt neutraalne |
| Purifying selection | puhastav selektsioon |
| Balancing selection | tasakaalustav selektsioon |
| Positive / negative selection | positiivne / negatiivne selektsioon |
| Directional selection | suunav valik |
| Stabilizing selection | stabiliseeriv valik |
| Disruptive selection | lõhestav valik |
| Overdominance | üledominantsus |
| English | Eesti keeles |
|---|---|
| Polymorphism | polümorfism |
| Silent polymorphism | vaikiv polümorfism |
| SNP | üksiknukleotiidne polümorfism |
| Allozyme | allosüüm, alloensüüm |
| Synonymous substitution | sünonüümne asendus |
| Non-synonymous substitution | mittesünonüümne asendus |
| Functional constraints | funktsionaalsed piirangud |
| Molecular clock | molekulaarne kell |
| Pseudogene | pseudogeen |
| English | Eesti keeles |
|---|---|
| Adaptation | kohastumus |
| Acclimation | aklimatiseerumine |
| Phenotypic plasticity | fenotüübiline plastilisus |
| Quantitative trait | kvantitatiivne tunnus |
| Heritability | päritavus |
| Additive genetic variance | aditiivne geneetiline varieeruvus |
| Inbreeding depression | inbriidingdepressioon |
| Cline | kliin |
| Gene expression | geeniekspressioon |
Because new neutral mutations each have fixation probability , and the two cancel to give . Under Ohta’s nearly neutral theory the class of mutations that behaves neutrally is set by , so decides how much of the genome is effectively neutral — and therefore the rate.
means more amino-acid polymorphism than expected from the divergence. Either purifying selection is removing slightly deleterious amino-acid variants that are still segregating, or the population has a small — a post-glacial Baltic population is a good candidate — so slightly deleterious variants drift to appreciable frequency. A recent bottleneck can produce the same signal without any change in selection.
No. Neutral markers measure gene flow and drift, and a handful of migrants per generation is enough to erase while leaving the populations adaptively distinct. Local adaptation can persist at a few loci under selection against a background of homogenised neutral variation — which is exactly why management should look at adaptive variation too.
A common-garden experiment: raise both populations for at least one full generation under identical conditions, then measure heat tolerance. Any difference that remains is genetic; what disappears was plasticity. A reciprocal transplant adds the environmental half of the answer, and eQTL mapping tells you which loci carry the expression response.
Tähtis
peeter.laas@taltech.ee · peeter.laas@ut.ee
Eestikeelne materjal:
A. Heinaru, Geneetika, evolutsioonigeneetika peatükk.
Seminarid: 15 min ettekanne + 5 min küsimusi. Ülesehitus: sissejuhatus (taust ja eesmärgid), meetodid, peamised tulemused, järeldused. Ole valiv — kõiki tulemusi ei pea esitama.
Fall 2026