INFERRING PROCESS FROM PATTERN

The mechanics of population genetics attempts to describe how alleles and genotypes change from one generation to the next. Often evolutionary biologists cannot study these transitions in real time, either because the generation time is too long (redwoods, elephants) or the species are fossils or other necessary information cannot be measured. In some cases the patterns of variation can suggest or indicate what processes have been acting. Ideally we would want to look at the patterns of variation and be able to say unequivocally that one or another process had produced that pattern. This is not always possible since identical patterns can be produced by very different processes. BUT: most patterns suggest obvious experiments that might tease apart various processes contributing to those patterns.

Example: trajectory of allele frequency change shows a gradual but rough increase to fixation. If we had this information alone we could not determine whether drift or selection drove the allele to fixation. Design testable hypothesis to distinguish possible causes: measure effective population size if pq/2Ne << p each generation might conclude selection drove allele frequency change.

Biston betularia: pattern suggests selection; what is hypothesis and how would you test it? Bird predation test: mark-release-recapture experiment in different areas: more dark preyed upon in rural areas (see figure 5.5, pg. 108), more light form preyed upon in industrial areas. Patterns of frequencies of dark form is very different in moths with different population structures. Biston betularia has low density and flies far to find mates: frequency of dark form is quite uniform in industrial areas. Gonodontis bidentata has higher density and is more localized: more variation in frequency of melanic forms. Different balance of gene flow and selection. Two points: 1) increased migration rate (m) leads to more homogenization of allele frequencies, 2) increases effective population size (N_e) means that selection will be more effective (weaker drift).

Heavy metal tolerance in plants. Selection inferred: mine populations have evolved resistance to high concentrations of copper, zinc and lead. Test: for selection of genetic basis: transplant plants across boundary and the "pasture" form cannot grow on the heavy metal soil: implies true genetic response. Pattern of distribution: cline in frequency of resistant forms as one moves across boundary. Suggests gene flow between tolerant and non-tolerant forms, but despite gene flow divergence is maintained: must be strong selection to overcome effects of gene flow. Examine patterns carefully: both the presence of a cline (frequency transition) and the "steepness" of the cline can say thing about process.

Shell polymorphism in Cepaea nemoralis: find different colors (pink, yellow, brown) and different banding pattern (five banded, two-banded, unbanded, etc.). Genetics worked out pretty well: one locus for color, another for banding number and other modifier loci. Pattern: find different morphs in different habitats. Is it selection or not? Hypothesis? = predation. Test=? look at so called "thrush-anvils" stones where the song thrush who preys on the snails breaks them open to eat. Look at the frequency of types of shells around thrush anvils and compare these frequencies to the frequencies in the general population. Apparent that there is differential predation by habitat: pink and unbanded snails more common in woods; yellow and banded snails more common in fields and hedgerows. Implies that visual predation is important character.

Allele frequency clines: Alcohol dehydrogenase locus (Adh) and a-glycero-phosphate dehydrogenase locus (a-GPDH) have cline along the east coast of N. America. Hypothesis=? selection differs with latitude. Test=? compare southern hemisphere: get same cline in opposite direction implies selection but difficult to identify a specific mechanism

Leucine amino peptidase locus (Lap) in Mytilus edulis polymorphism varies with habitat and position in Long Island Sound Lap94 allele found in higher salinity waters. Hypothesis=? salinity. Test=? sample at different time of year with different runoff; find predicted results. Physiology: determine how the different alleles work in the cell and differ in performance. Results lead to a prediction of which should be in certain environments. Test prediction with additional sampling. Test model of action of selection: phenotype could be additive, mutliplicative or dominance. Each genetic model gives different fitnesses, hence predicts different deviation from Hardy Weinberg genotype frequencies. Data fit predictions of one model (dominance) well.