How frequent is adaptation? Does it generally involve mutations of small or large phenotypic effect? Does it tend to involve new mutations or use standing variation? Are adaptive mutations typically coding or regulatory? How frequent are population or environment-specific versus whole-species adaptations? What effect does recurrent adaptation has on patterns of neutral variation? Is the Neutral Theory dead?

Quantifying molecular adaptation

We attempt to quantify the rate and the strength of adaptation from genome-wide patterns of neutral polymorphism and functional divergence (51).

Fig. The pattern of synonymous polymorphism along the chromosome 2L in D. simulans. The high levels of variability might be partially due to persistent adaptation.

We estimate that adaptation is both very frequent (approximately one every 1000 generations) and strong (on the order of 1% advantage). If true, this would mean that virtually all neutral polymorphisms in the Drosophila genome are affected by adaptations in their genomic vicinity. In collaboration with Guy Sella at Hebrew University we are conducting similar analyses in humans and yeast.

Adaptation due to transposable elements

Transposable elements (TEs) are powerful mutagens. In Drosophila transposable elements might generate the majority of visible mutations. We have systematically searched for adaptations generated by TEs in D. melanogaster. We have collected population frequency data for a large fraction of transposable elements found in the sequenced D. melanogaster genome and searched for those TEs that have sharply increased in frequency in the recent past. Our search uncovered 13 putatively adaptive TEs implying a very high rate of TE-induced adaptation (54). We are currently investigating the population genetics and phenotypic effects of these TE insertions.

Evolution of pesticide resistance

We invetigate evolution of pesticide resistance and of other adaptations to human alterations in the environment. We have discovered that a disruption of a conserved choline kinase by a TE (38) in D. melanogaster leads to pesticide resistance. We also investigate the population genetics and phenotypic effects of resistance mutations in the acetylcholine esterase gene (Ace). We also are taking a genome-wide approach to finding very recent adaptations in D. melanogaster.

Adaptation to fresh-water environment in three-spine sticklebacks

We are collaborating with the laboratory of David Kingsley and CEEG to study the evolution of adaptation in three-spine sticklebacks. Our laboratory specifically focused on the inference of the demographic events that were involved in the colonization of fresh-water habitats by three-spine sticklebacks. We also investigate signatures of selection around loci that underlie adaptations in the fresh-water populations of sticklebacks.

Demography and non-adaptive explanations for signatures of selective sweeps

In order to study adaptation we investigate the alternative processes that can generate signatures resembling those left by adaptation (such as selective sweeps). Demographic perturbations (such as bottlenecks) are notorious is generating such signatures. We also investigate additional possibilities (such as specific ascertainment biases, increase in frequency of slightly deleterious variants, and restrictions of recombination in heterozygotes around DNA insertions) (52). In general, this line of research is aimed at refining our ability to detect true adaptations. Here are some of the publications on this topic (50, 52, 55).


Why are some genomes small and others large? What determines the numbers and kinds of genes found in different genomes? Chromosomal numbers? How do multigene families come to be and maintain their sizes? What role does horizontal gene transfer have on genome evolution? Mutational biases? Natural selection?

Evolution of point mutation patterns

Fig. Patterns of background substitution rater on Chromosome 1 in humans

A key parameter in molecular evolution is the background pattern of substitution -- the pattern of substitution that we would observe in the absence of selection for information-coding functions of the kind experienced by genes. Normally this task is accomplished through the use of pseudogenes -- nonfunctional copies of functional genes. However, it is often difficult to find enough pseudogenes, especially in poorly studied or compact, pseudogene-poor genomes. We have devised an alternative method that uses the unconstrained evolution of defunct, "dead" transposable elements ("deadTE" approach) (08). Using this approach we quantified patterns of background point substitution in Drosophila (14, 34, 46) and in mammals (29, 37).

Evolution of genome size

We have also used the deadTE approach to quantify rates and patterns of small insertions and deletions (indels) in a number of organisms (reviewed in 21 and 22). We discovered that eukaryotes that have very compact genomes tend to have much high rate of DNA loss via indels than organisms with large genomes. We investigate the role that variation in indel biases play in generation of a staggering, 200000-fold variation in genome size among eukaryotes. We also investigate the underlying dynamics (48) and adaptive significance (33) of genome size evolution.

Evolution of TEs

We investigate the forces that limit the spread of TEs in Drosophila and mammals. We have provided evidence that supports the notion that selection against the deleterious effects of ectopic recombination is a dominant force in preventing many TE families in Drosophila from increasing in copy number (27). We also provided evidence that the same process is operating in the human population generating genetic load due to a large number of TE copies in the human genome (45).

Evolution by gene loss and gene duplication

We are interested in why some genes regularly generate duplicate copies and others do it very rarely. We have provided evidence that genes that produce persistent duplicates tend to evolve slowly (30). We are carrying out additional computational work trying to understand other determinants of duplicability.

We are also interested in the process by which duplicate genes persist in the genome. In collaboration with the laboratory of Ron Davis, we have carried out a large-scale deletion analysis of pairs of duplicated genes revealing that redundancy is widespread among even very old duplicated genes and that duplicated genes do not appear to evolve much additional functionality (at least in the rich medium) (53).

Finally we investigate the determinants of gene persistence in the genome. We have evidence that persistent genes tend to be functionally important and that such genes also tend to underlie adaptation less frequently (56).