Quantitative Evolutionary Genetics Drosophila in the Post-Genome Era
Symposium organized by Jean David and Volker Loeschcke from 24 – 29 April 2001
Anholt Robert (North Carolina State University, Raleigh, USA), Bijlsma Kuke Bureau (University of Groningen, Haren, Pays-Bas), Jean-François Capу (Institut Pasteur, Paris CNRS, Gif-sur-Yvette CNRS) Pierre David Jean R. De Long (University of Utrecht, Pays-Bas), Gerdien Gibert (CNRS, Gif-sur-Yvette), Patricia Gilchrist (University of Sydney, Australie), Stuart Hoffmann (La Trobe University, Victoria, Australie), Ary Ioule David (Florida State University, Tallahassee, USA), Joshi Amitabh (Jarvaharlal Nehru Centre for Advanced Sc. Research, Jakkuv Bangalore, Inde), Klingenberg Christian (University Museum of Zoоlоgy, Cambridge,Grande‑Bretagne), Loeschcke Volker (Aarhus University, Aarhus, Danemark), Mac Kау Trudy (North Carolina State University, Raleigh, USA), Moreteau Brigitte (CNRS, Gif-sur-Yvette), Mott Richard (Wellcome Trust Centre for Humas Genetics, Oxford, Grande-Bretagne), Tixier-Boichard Michèle (Institut national de la recherche agronomique, Jouy-en-Josas), Van de Zande Louis, (University of Groningen, Haren, Pays-Вas), White Kevin (Yale University, New Haven, USA).
1. General aims of the Conference
Ever since Darwin, natural selection, acting on phenotypic variation, has been considered as the deterministic mechanism of biological evolution. White a major feature of Darwinian theory is gradualism, most genetic analyses of individual variability have focused on discontinuous traits. Population genetics, considering variations of allelic frequencies in nature, is a classical approach for bridging the gap between genetics and evolution. Numerous people (e.g. R.C. Lewontin) have however pointed out that population genetics, using theoretical models or molecular data, provides all needed information on population variability and polymorphisms, but very little on adaptation. The major challenge still remains to bridge the gap between genotype aid phenotype, and one way to do so is quantitative genetics.
Up to now, quantitative genetics has mainly been a statistical description of phenotypes. The classical genetic interpretation assumes the occurrence of numerous genes with small aid additive effects, called polygеnes. Unfortunatеly, the existence of real polygones is becoming unlikеlу, in the light of progress made by molecular and developmental genetics: genes acting on quantitative traits (QTL) are beginning to be identified; manу genes are highly pleiotropic; regulatory sequences are certainly involved in evolution; interactions between genes (epistas) are numerous and multiform.
Quantitative Evolutionary Genetics (QEG)
The statistical approach has been very successful for applied quantitative genetics. Improving crop or animal production does not require the genetic bases of a trait of interest to be precisely known.
In the field of evolutionary biology, major progress has been made over the last two decades in our understanding of life history evolution. Fitness traits such as viability, developmental rate, body size and age at maturity have been compared between species and their evolution modelized. The occurrence of multiple interactions aid of negative correlations (trade-offs) appears as a general trend accounting for the diversity of extant species. However the genetic bases of such traits remain basically unknown, and QEG now appears as the new approach for a better understanding of evolutionary mechanisms. More preсisely, evolutionists are not concerned with the artificial improvement of certain performances, but want to work out the genetic bases of life history diversity.
Drosophila as an ideal model
Drosophila is alrеadу a model in evolutionary biology, with more than 3000 species described in the familу. These have a world-wide distribution, a diversity of ecological niches, well documented life histories, adaptations and phylogenies. The fact that several hundreds of species can be reared in the laboratory and are amenable to experimental and genetical analyses is an irreplaceable advantage.
Within this huge set of species, Drosophila melanogaster appears as an ideal starting point for understanding the genetic bases of numerous adaptations and life history traits, for three major reasons:
— a large amount of information already exists, concerning morphology, development, physiology and behaviour. A precise genetic map is available and also special techniques such as balancer chromosomes, P element mutagenesis and polytene chromosomes.
- the species exists in a diversity of geographic races (e. g. tropical and temperate) which reveal adaptive genetic divergence occurring in a recent
- the year 2000 provided an almost complete genomic sequence, with numerous genes of unknown function, offering a starting point for future analyses. DNA microarrays should be extremely useful in this respect.
The proposal was that all participants should have an interest in quantitative variations and quantitative genetics, most but not ail in QEG, and most but not all in Drosophila.
The general objective was to give a state of the art overview of the discipline and to focus also on prospective researches. Among numerous problems which were expected to be debated for improving the future, we may mention: working out new directions of research; identifying possible pitfalls (theoretical and practical); defining techniques and guidelines; discussing the relative merits of top-down and bottom-up approaches; considering methods appropriate to investigate gene regulation; considering the value of the Drosophila model for human genetic analyses and understanding gene interactions; canalising future researches into the most promising directions.
2. Report on the meeting and scientific conclusions
The workshop was hold in April 2001 and lasted 4 days. The 19 participants, from various parts of the world, gave 19 talks of about 50 minutes, and each talk was followed by an active discussion. Among the 19 participants 16 had a direct experience of Drosophila research, and the 3 others used mice (2) or chicken (1) as a model organism. Besides these presentations, general discussions were organised, on a diversity of relevant themes, for a total duration of more than 3 hours. The list of these themes is given in Annexe I while the conclusions are incorporated below. The summaries of the oral presentations are given in Annexe II.
Being an integrative discipline, Evolutionary Genetics requires all kinds of information on the Drosophila model. Communications ranged from the eсology of natural populations, trying to understand natural selection and phenotypic adaptation, to the utilisation of molecular techniques of functional genomics, trying to work out the genetic bases of adaptive quantitative traits.
Drosophila natural populations mау live in different climates and often exhibit latitudinal clines which are considered as local climatic adaptations (to temperature and humidity). Original comprehensive results were presented for the East coast populations of two species in Australia. Adaptations were described for body size, starvation tolerance, resistance to temperature extremes (A. Hofmann) . External conditions are often stressful, at least during part of the dау or of the year. One side of thermal adaptation is heat tolerance, which is related to the induction of HSP proteins by heatshocks. A detailed analysis of the interactions between stresses, HSP induction and genetic background on a diversity of fitness traits was presented, demonstrating the complexity of ecological adaptations in two Drosophila species (V. Loeschcke).
Within a given population, inbreeding reveals a large amount of genetic load which is a means for studying QTL, and also a model for conservation biology (R. Bijlsma). Crosses between geographically distant populations are used for analysing the genetic architecture of quantitative traits. Complex genetic interactions, including an outbreeding depression, evidenced the genomic integration of local populations (S. Gilchrist). Fitness-related traits and life history evolution were considered in several talks. Alcohol adaptation related to alcohol dehydrogenasе activity is a classical trait in D. melanogaster. A sophisticated molecular analysis of selected laboratory lines revealed some significant differences independent of the Adh gene, difficult however to interpret in terms of gene action and regulation (L. van de Zande). Rate of larval development was analysed in several talks, analysing either the possible responses to selection (G. dе Jong and A. Joshi) or the genes expressed in different stages (K. White). These complementary approaches revealed the major importance of the developmental rate among life history traits, and the possibility to identify the interactions between metabolic activity and hormonal secretions. Longеvity is expected to be a complex trait in natural populations.
This was clearly shown by a QTL mapping analysis, which revealed a large number of gene-environment interactions and sometimes unexpected differences between sexes (T. Mackay). Behaviour genetics is generally a difficult discipline. Simple olfactory tests permitted the identification of QTL and then the identification of several candidate genes involved in olfactory networks (R. Anholt). Complex traits can be amenable to genetic analysis; one is phеnotypic plasticity, another is wing shape. Phenotypic plasticity is the property by which a given genotype maуproduce different phenotypes in different environments. The shape of the response curve (the reaction norm) can be analysed by specific methods. The major conclusion is that plasticity is a specific property, independent of trait value and which ray be submitted to natural selection. The occurrence of specific plasticity genes remains, however, controversial (P. Guibert). Wing size (length or area) is certainly an adaptive trait, while wing shape (the architecture of veins) needs more complex investigation and interpretation, which were discussed in three presentations. Phenotypic plasticity in two related species was a means to reveal) a complex genetic architecture. Classical taxonomic indices, used to get rid of size variations, were also shown to be plastic (B. Moreteau). The two other talks used mainly multivariate analyses. Procrustes methods were used to describe changes in the position of specific landmarks, permitting the description of the effects of genetic changes (related to known mutants) and of developmental instability (left-right-side asymmetry) (C. Кingenberg).
The major problem of dimensionality (the least number of dimensions which are needed to describe variation at the phenotypic and genetic levels) was addressed comparing wing shapes in different species. A new technique, called the minimum dimensional space, was presented, permitting comparisons at several levels of the evolutionary process (D. Houle). Only two presentations considered quantitative variations between species. Regulation of gene expression can be analysed with DNA microarrays (K. White). Classical quantitative traits mау also be investigated in interspecific crosses, and sometimes unexpected results reveal complex interactions. Hybrid breakdown occurs in certain traits but not in others, and introgressions are difficult due to genomic incompatibilities (J. David). The evоlutionary significance of transposable elements was considered at various levels of evolution: local populations, closely related species, distantly related species or distant taxa. In several cases, their role in generating a new genetic variance appears clearly, and also their importance in genome evolution (P. Caру).
Phenotypic variations were analysed in different ways at the gene level. Genetics of pigmentation variation in chicken and mice is a means of revealing the functional interactions of known developmental genes. Pigmentation mutants are homologous among birds and mammals and genеrally imрlу qualitative changes at the protein level (M. Tixier Boichard). One presentation dealing with the mouse model looked at the technical and theoretical problems which are involved in the QTL detection. A possible analysis starting from a multiple cross followed by several generations of recombination was presented (R. Mott). The susceptibility/tolerance of mice to a neuronal picornavirus was analysed in various strains. There is the major role of the H-2 region, but also the significant effects of other major genes on other chromosomes. Also the various phenotypes of the viral infection respond to different genes, revealing the сomplexity of ail epistatic interactions (J.F. Bureau). In Drosophila, bristle number, related to the expression of neuronal gene, is a most investigated trait. A progressive analysis of QTL was presented, using different resolution techniques such as recombinant inbred lines, deficiency complementation mapping, P element mutagenesis. Also candidate pleiotropic genes were known from developmental genetics. When the analysis reaches the DNA level, QTL are often explained bу changes in rеgulatory regions and not by protein differences (T Мackay).
3. Conclusions and Prospects
A first general conclusion is that it is necessary, and also possible to bridge the gap between scientific disciplines which up to now have evolved indepеndently, that is quantitative genetics, developmental genetics and ecological genetics.
The ultimate goal is to understand the genetic basis of complex phenotypiс traits, related to fitness and adaptation in natural populations. This will be possible in a few model species, such as Drosophila, mice and humans, which are amenable e to genetic analyses and for which a complete sequence of the genome is known. Two simultaneous approaches are possible and needed: the bottom-up approach, using functional genomic DNA microarrays or proteomic studies; the top-down approach, trying to define traits of interest in natural population (QTL) and progressively working out their genetic determinism, ultimately at the DNA level.
The scientific community must become aware that gene interactions and regulations are the major basis of quantitative variation. Such problems are exceedingly complex but may be solved, at least in some favourable cases, by some kinds of functional genomics. It is always to be reminded that similar phenotypes performing the same task can be determined by different genotypes within the same species. The broad diversity of extant available techniques and concepts can and must be used to resolve two major unrelated challenges: unravelling the genetic mechanisms leading to the origin of species, and especially the genomic incompatibilities among populations; understanding the genetic bases of complex quantitative diseases in human populations at an individual level.
Annexe 1: Scientific themes discussed in general meeting
- What maintains the genetic variance in quantitative traits?
- How many (which) genes are involved in adaptation? How manу genes are involved in variation within populations? How manу genes are involved in speciation?
- What is the relationship between gene networks and statistical descriptions of phenotypic and genetic variation? Are there regularities in genotype-phenotype caps that can help to generalize the network/statistical relationship?
- Does evolutionary change occur more at regulatory or at structural sites? How can we measure “more”?
- Is there a difference in the rate of evolution of developmental vs. ecological traits? Early vs. late? Life history vs. morphological traits?
- Are alternative paths to the same adaptation often equivalent in fitness?
- How prevalent are epistatic interactions at the molecular, physiological, phenotypic level? Are molecular interactions more likely to occur at transcription level?
- What is the dimensionality of phenotypic variation?
- To what extent do laboratory estimates of quantitative genetic parameters inform us about evolutionary patterns in natural populations?
- What do fitness landscapes look like?
- What do quantitative, whole organism biologists have to offer the postish-genomic world?
- How do the new genomic techniques help answer the old questions?
- What is the relationship between species variation and among species variation? Is classical quantitative genetics at al relevant to the study of adaptation? Are the same loci involved? Are the same alleles involved?
- What limits evolutionary rate? Populations structure? Mutation or variation in populations?
- The credibility of quantitative genetics. Towards establishing data bases of a given species on a world-wide basis.