We have found that constant selection against mutations can cause cyclical dynamics in a population with facultative selfing. When this happens, the distribution of the number of deleterious mutations per genotype fluctuates with the period ∼1/sHe generations, where sHe is the coefficient of selection against a heterozygous mutation. The amplitude of oscillations of the mean population fitness often exceeds an order of magnitude. Cyclical dynamics can occur under intermediate selfing rates if selection against heterozygous mutations is weak and selection against homozygous mutations is much stronger. Cycling is possible without epistasis or with diminishing-returns epistasis, but not with synergistic epistasis. Under multiplicative selection, cycling might happen if the haploid mutation rate exceeds 1·9 in the case of selfing of haploids, and if this diploid mutation rate exceeds 4·5 in the case of selfing of diploids. We propose a heuristic explanation for cycling under facultative selfing and discuss its possible relevance.

Intrapopulation variability in the seasonal regulation of insect lifecycles has been shown to be due partly to genetic changes. Selection for insecticide resistance in the codling moth Cydia pomonella results from allelic substitution at two to three loci in south-eastern French populations of this species. However, such an adaptive process has been associated with an increased heterogeneity in the developmental responses to climatic factors such as temperature. In this paper, we investigate whether such pleiotropic effects of resistance on development induce a significant discrepancy in seasonal regulation in this species. The seasonal changes in a susceptible and two insecticide-resistant homozygous genotypes of C. pomonella, as well as their reciprocal F1 progeny, were followed under natural conditions during the reproductive season through the emergence events of adults, within-generation developmental rates and the number of generations. A significant delay in the occurrences of homozygous resistant genotypes resulted from significantly lower pre-imaginal developmental times relative to homozygous susceptible ones. Subsequent assessment of the number of generations indicated significantly higher diapause propensities in carriers of the resistance alleles (37·0–76·2%) than in susceptible homozygotes (6·7%), which mostly pupated towards a third generation of adults. In the light of these findings, pleiotropic effects of adaptive changes might be a crucial source of divergence in seasonal regulation at the population level, involving significant life-history trade-offs. In addition to man-made selective factors during the reproductive season, such an effect on the lifecycle could be a key component in the process of selection for resistance genes in south-eastern France C. pomonella populations.

Recombination is thought to have various evolutionary effects on genome evolution. In this study, we investigated the relationship between the base composition and recombination rate in the Drosophila melanogaster genome. Because of a current debate about the accuracy of the estimates of recombination rate in Drosophila, we used eight different measures of recombination rate from recent work. We confirmed that the G+C content of large introns and flanking regions is positively correlated with recombination rate, suggesting that recombination has a neutral effect on base composition in Drosophila. We also confirmed that this neutral effect of recombination is the main determinant of the correlation between synonymous codon usage bias and recombination rate in Drosophila.

The usage of preferred codons in Drosophila melanogaster is reduced in regions of lower recombination. This is consistent with population genetics theory, whereby the effectiveness of selection on multiple targets is limited by stochastic effects caused by linkage. However, because the selectively preferred codons in D. melanogaster end in C or G, it has been argued that base-composition-biasing effects of recombination can account for the observed relationship between preferred codon usage and recombination rate (Marais et al., 2003). Here, we show that the correlation between base composition (of protein-coding and intron regions) and recombination rate holds only for lower values of the latter. This is consistent with a Hill–Robertson interference model and does not support a model whereby the entire effect of recombination on codon usage can be attributed to its potential role in generating compositional bias.

222 cultivated (Vitis vinifera) and 22 wild (V. vinifera ssp. sylvestris) grape accessions were analysed for genetic diversity and differentiation at eight microsatellite loci. A total of 94 alleles were detected, with extensive polymorphism among the accessions. Multivariate relationships among accessions revealed 16 genetic groups structured into three clusters, supporting the classical eco-geographic grouping of grape cultivars: occidentalis, pontica and orientalis. French cultivars appeared to be distinct and showed close affinity to the wild progenitor, ssp. sylvestris from south-western France (Pyrenees) and Tunisia, probably reflecting the origin and domestication history of many of the old wine cultivars from France. There was appreciable level of differentiation between table and wine grape cultivars, and the Muscat types were somewhat distinct within the wine grapes. Contingency χ2 analysis indicated significant heterogeneity in allele frequencies among groups at all loci. The observed heterozygosities for different groups ranged from 0·625 to 0·9 with an overall average of 0·771. Genetic relationships among groups suggested hierarchical differentiation within cultivated grape. The gene diversity analysis indicated narrow divergence among groups and that most variation was found within groups (∼85%). Partitioning of diversity suggested that the remaining variation is somewhat structured hierarchically at different levels of differentiation. The overall organization of genetic diversity suggests that the germplasm of cultivated grape represents a single complex gene pool and that its structure is determined by strong artificial selection and a vegetative mode of reproduction.

A composite cross was made between 12 strains of the fungus Ascobolus immersus, six with wild-type red ascospores (w1+) and six with white ascospore mutation w1-78. A high postmeiotic segregation (PMS) frequency line was set up from colonies from ascospores from dehisced octads showing PMS, 5+[ratio ]3w and 3+[ratio ]5w. A low PMS line was started from ascospores from 4+[ratio ]4w or 6+[ratio ]2w octads, and a ‘no selection’ line was set up from ascospores from random octads. Colonies were crossed to tester strains to determine PMS frequencies and the selected lines were continued from ascospores of crosses of the red ascospore strain with the most extreme (e.g. high for the high line) PMS frequency with the white-ascospore strain of most extreme PMS frequency and of opposite mating type. Significant responses to selection were obtained for increased (+100%) and decreased (−58%) PMS, giving a 4·8-times difference in generation 4, with little change in the frequencies of conversion classes showing meiotic segregation (6+[ratio ]2w and 2+[ratio ]6w). The continuous, symmetrical, roughly normal distributions for PMS frequencies obtained when generation 5 strains were crossed to unselected tester strains are those expected if PMS frequencies are controlled by a number of polygenes, not major genes. Crosses of selected fifth-generation red-ascospore strains with extreme PMS values to base-substitution mutant w1-78, to frame-shift mutant w1-3C1 and to white-ascospore mutants w-BHj and w-9 at two loci unlinked to w1 showed that the effects of selection were not allele specific, locus specific or mutation-type specific.

The soil nematode Caenorhabditis elegans is an example of a species in which self-fertilizing hermaphrodites predominate, but functional males continue to persist – allowing outcrossing to persevere at low levels. Hermaphrodites can produce male progeny as a consequence of sex chromosome non-disjunction or via outcrossing with males. Consequently, the genetics of sex determination coupled with the efficiency by which males find, inseminate and obtain fertilizations with hermaphrodites will influence the frequency at which males and outcrossing occurs in such populations. Behavioural and physiological traits with a heritable basis, as well as ecological characters, may influence male reproductive success and therefore sex ratio. Because sex ratio is tied to male reproductive success, sex ratio greatly affects outcrossing rates, patterns of genetic variation, and the ability of natural selection to act within populations. In this paper we explore the determinants of male frequency in C. elegans with a mathematical model and experimental data. We address the role of the genetic machinery of sex determination via sex chromosome non-disjunction on sex ratio and the influence of physiological components of C. elegans' life history that contribute to variation in sex ratio by way of male reproductive success. Finally, we discuss the short-term and long-term factors that are likely to affect sex ratio and breeding system evolution in species like C. elegans.

Several forces may affect the distribution of genetic diversity in natural populations when compared to what is expected in a random-mating, constant size population of neutral genes. One solution for unravelling their respective influence is to study several genes at once in order to better reflect the true genealogy. Here we reconstruct the evolutionary history of the freshwater snail Biomphalaria pfeifferi over its entire distribution, using eight African populations, and three congeneric species as outgroups. A phylogenetic analysis was conducted using amplified fragment length polymorphism markers, and sequences at eight nuclear non-coding loci and one mitochondrial gene were used to analyse population structure. The geographic distribution of variation suggests greater affinities within than among regions. The pattern of variability at both the nuclear and mitochondrial DNA (mtDNA) loci is consistent with a bottleneck, although population structure may also partly explain our results. Our results are also indicative of the role of selection, whether positive or purifying, in the mtDNA. This highlights the fact that the interfering influences of population structure, demography and selection on molecular variation are not easily distinguished.

Genetic and cytological approaches have yielded significant insight into the mapping and organization of genes located in the heterochromatin of Drosophila melanogaster. To date, only a few of these genes have been molecularly characterized in detail, and their function unveiled. As a further step towards the identification of heterochromatic gene functions, we have carried out a cytological analysis of mitotic and meiotic cell divisions in mutants carrying different allelic combinations of l(2)41Aa, a gene located in the proximal heterochromatin of chromosome 2. Our results showed that larval brains of l(2)41Aa mutants display a high frequency of cells with irregularly condensed chromosomes. In addition, defective chromosome condensation was detected in male meiosis, consequently affecting chromosome segregation and giving rise to irregular spermatids. Taken together, these findings indicate that l(2)41Aa is a novel cell cycle gene required for proper chromosome condensation in both somatic and germ line cells.

We investigated the phenotypic plasticity of sternopleural bristle (SB) number as a function of growth temperature in isofemale lines from temperate (France) and tropical (Congo) populations of Drosophila melanogaster. We found concave reaction norms with a maximum in the middle of the thermal range, except in four African lines which exhibited a regularly decreasing response curve. Genetic variability (intraclass correlation) and evolvability (genetic CV, coefficient of variation) were independent properties and did not change with temperature. Residual, within-line variability was, however, strongly influenced by growth temperature, showing a U-shaped response curve and a minimum CV of 9% at 21·5 °C. As expected from a previously known latitudinal cline, maximum values (MV) were higher in temperate than in tropical flies. The temperature of maximum value (TMV) was observed at a higher temperature in the tropical population, in agreement with similar adaptive trends already observed for other quantitative traits. Significant negative correlations within each population were observed between a plasticity curvature parameter and MV or TMV. No difference in curvature was, however, observed between populations, in spite of their very different MVs.

No exact method for determining genotypic and identity-by-descent probabilities is available for large complex pedigrees. Approximate methods for such pedigrees cannot be guaranteed to be unbiased. A new method is proposed that uses the Metropolis–Hastings algorithm to sample a Markov chain of descent graphs which fit the pedigree and known genotypes. Unknown genotypes are determined from each descent graph. Genotypic probabilities are estimated as their means. The algorithm is shown to be unbiased for small complex pedigrees and feasible and consistent for moderately large complex pedigrees.

As part of a population genetics survey of the hybrid zone between mouse subspecies Mus musculus domesticus and M. m. musculus, we identified and characterized the t haplotypes in 1068 mice from 186 different populations in a 2500 km2 area in central Jutland. On the basis of two t-specific PCR markers, 130 mice possessed this haplotype. The allele frequencies at six microsatellites on the third and fourth chromosomal inversions of the t region were sufficiently different between t-bearing and non-t-bearing mice, and linkage disequilibria sufficiently marked on the t haplotype, to be able to reconstitute the genotype of most t haplotypes. A total of three frequent and 15 rarer haplotypes were identified. These haplotypes resemble each other more than they resemble a panel of known haplotypes from a wide range of geographical regions, except for tw73, which was also extracted from Jutland. The patterns of variation at the microsatellite loci suggest that the Jutland haplotypes were derived from a small number of haplotypes, followed by recombination between complementing haplotypes. Further evidence of recombination came from complementation tests that we performed, showing the lack of concordance between the degrees of complementation and of molecular resemblance between haplotypes. This study shows that it is possible to characterize the presence and variation of t haplotypes by a population genetics approach using simple molecular markers. However recombination between t haplotypes has occurred frequently enough to obscure the links between this variation and the biological properties of distortion and lethality of the haplotypes that originally colonized Jutland.

Current sitewise methods for detecting positive selection on gene sequences (the de facto standard being the CODEML method (Yang et al., 2000)) assume no recombination. This paper presents simulation results indicating that violation of this assumption can lead to false positive detection of sites undergoing positive selection. Through the use of population-scaled mutation and recombination rates, simulations can be performed that permit the generation of appropriate null distributions corresponding to neutral expectations in the presence of recombination, thereby allowing for a more accurate estimation of positive selection.

In Drosophila, sperm length and the length of the females' primary sperm-storage organ have rapidly coevolved through post-copulatory sexual selection. This pattern is evident even among geographic populations of Drosophila mojavensis. To understand better these traits of potential importance for speciation, we performed quantitative genetic analysis of both seminal receptacle length and sperm length in two divergent populations. Parental strains, F1, F1 reciprocal (F1r), F2, F2r, backcross and backcross reciprocal generations were used in a line-cross (generation means) analysis. Seminal receptacle length is largely an autosomal additive trait, whereas additivity, dominance and epistasis all contributed to the means of sperm length. Either an X-chromosome or a Y-chromosome effect was necessary for models of sperm length to be significant. However, the overall contributions from the X and Y chromosomes to sperm length was small.

Chromosomal rearrangements such as Robertsonian (Rb) fusions constitute a major phenomenon in the evolution of genome organization in a wide range of organisms. Although proximate mechanisms for the formation of Rb fusion are now well identified, the evolutionary forces that drive chromosomal evolution remain poorly understood. In the house mouse, numerous chromosomal races occur in nature, each defined by a unique combination of Rb fusions. Among the 106 different Rb fusions that were reported from natural populations, the low involvement of chromosome 19 in Rb fusions is striking, prompting the question of the randomness of chromosomal involvement in Rb fusions. We uncover a significant quadratic relationship between chromosome size and probability of fusing, which has never previously been in this species. It appears that fusions involving chromosome 19 are not particularly infrequent, given the expected low fusion probability associated with the chromosome's size. The results are discussed, assessing selective processes or constraints that may operate on chromosome size.

An offspring genome can be viewed as a mosaic of chromosomal segments or haplotypes contributed by multiple founders in any quantitative trait locus (QTL) detection study but tracing these is especially complex to achieve for outbred pedigrees. QTL haplotypes can be traced from offspring back to individual founders in outbred pedigrees by combining founder-origin probabilities with fully informative flanking markers. This haplotypic method was illustrated for QTL detection using a three-generation pedigree for a woody perennial plant, Pinus taeda L. Growth rate was estimated using height measurements from ages 2 to 10 years. Using simulated and actual datasets, power of the experimental design was shown to be efficient for detecting QTLs of large effect. Using interval mapping and fully informative markers, a large QTL accounting for 11·3% of the phenotypic variance in the growth rate was detected. This same QTL was expressed at all ages for height, accounting for 7·9–12·2% of the phenotypic variance. A mixed-model inheritance was more appropriate for describing genetic architecture of growth curves in P. taeda than a strictly polygenic model. The positive QTL haplotype was traced from the offspring to its contributing founder, GP3, then the haplotypic phase for GP3 was determined by assaying haploid megagametophytes. The positive QTL haplotype was a recombinant haplotype contributed by GP3. This study illustrates the combined power of fully informative flanking markers and founder origin probabilities for (1) estimating QTL haplotype magnitude, (2) tracing founder origin and (3) determining haplotypic transmission frequency.

Explaining the repeated evolution of similar sets of traits under similar environmental conditions is an important issue in evolutionary biology. The extreme alternative classes of explanations for correlated suites of traits are optimal adaptation and genetic constraint resulting from pleiotropy. Adaptive explanations presume that individual traits are free to evolve to their local optima and that convergent evolution represents particularly adaptive combinations of traits. Alternatively, if pleiotropy is strong and difficult to break, strong selection on one or a few particularly important characters would be expected to result in consistent correlated evolution of associated traits. If pleiotropy is common, we predict that the pattern of divergence among populations will consistently reflect the within-population genetic architecture. To test the idea that the multivariate life-history phenotype is largely a byproduct of strong selection on body size, we imposed divergent artificial selection on size at maturity upon two populations of the cladoceran Daphnia pulicaria, chosen on the basis of their extreme divergence in body size. Overall, the trajectory of divergence between the two natural populations did not differ from that predicted by the genetic architecture within each population. However, the pattern of correlated responses suggested the presence of strong pleiotropic constraints only for adult body size and not for other life-history traits. One trait, offspring size, appears to have evolved in a way different from that expected from the within-population genetic architecture and may be under stabilizing selection.

In QTL analysis of non-normally distributed phenotypes, non-parametric approaches have been proposed as an alternative to the use of parametric tests on mathematically transformed data. The non-parametric interval mapping test uses random ranking to deal with ties. Another approach is to assign to each tied individual the average of the tied ranks (midranks). This approach is implemented and compared to the random ranking approach in terms of statistical power and accuracy of the QTL position. Non-normal phenotypes such as bacteria counts showing high numbers of zeros are simulated (0–80% zeros). We show that, for low proportions of zeros, the power estimates are similar but, for high proportions of zeros, the midrank approach is superior to the random ranking approach. For example, with a QTL accounting for 8% of the total phenotypic variance, a gain from 8% to 11% of power can be obtained. Furthermore, the accuracy of the estimated QTL location is increased when using midranks. Therefore, if non-parametric interval mapping is chosen, the midrank approach should be preferred. This test might be especially relevant for the analysis of disease resistance phenotypes such as those observed when mapping QTLs for resistance to infectious diseases.

When multiple related families derived from inbred lines are jointly analysed to detect quantitative trait loci (QTLs), the analysis should estimate allelic effects as accurately as possible and estimate the probability that different parents carry alleles that are identical in state. Analyses exist that assume that all parents carry unique alleles or that all parents but one carry the same allele. In practice, many configurations are possible that group different parents according to their identity-in-state condition at a putative QTL allele. Here, we propose a variable model Bayesian analysis that selects among possible identity-in-state configurations and jointly estimates the allelic effects of identical-in-state parents. We contrast this analysis with a fixed model analysis that estimates unique allelic effects for all parents. We analyse two simulated mating designs: an experimental design in which three inbred parents were crossed to generate two families of 150 doubled haploid lines; and a breeding design in which 20 inbred parents were crossed to generate 60 families of 20 doubled haploid lines, with each parent contributing to six families. In all cases where some parents were simulated to carry alleles of identical effect (that is, they were identical in state), the variable analysis estimated allelic effects with lower mean-squared error than the fixed analysis. The variable analysis showed that, unless each family contains many individuals (more than 100), there is insufficient information in DNA-marker and phenotypic data to determine with high probability the QTL allelic number.