DFG FOR 1078 Natural Selection in Structured Populations

Breadcrumb Navigation


Results of first funding period

The key results we have reached so far are as follows:

  • we have an understanding of the history (including subdivision and demography) of diverse species, including Drosophila, ants, mice, and bacteria,
  • we have an understanding of the form of selection in natural populations of these populations and species, and
  • we have developed a suite of analytical and simulation tools for studying natural selection in populations with complex demography and substructure.

In particular, we have now a more complete grip on the demographic history of D. melanogaster by having analyzed additional populations by Approximate Bayesian Computation (ABC): a population from the ancestral species range (Africa) and two derived populations (Asia and Europe). Specifically, we have confirmed that there is only a single major postglacial migration route out of Africa (via the Near East). And by QTL and selective sweep analyses we have begun to identify the genes involved in adaptation to temperate zones.

Furthermore, expression studies have revealed a suite of new genes that are differentially expressed between African and European populations of D. melanogaster. The selective and functional causes of these differences at some of these genes have been further investigated.

In all major subspecies of house mice and in a more distantly related species (M. spretus), a clear pattern of trans-species polymorphism was found at the B4galnt2 locus. This indicates that the hypothesized long-term balanced polymorphism at the B4galnt2 locus conferring alternative tissue-specific expression patterns is more than two million years old.

In ants, the influence of population structure on the evolution of the defense trait slave rebellion has been hypothesized to be important: the destruction of parasite brood increases the inclusive fitness of enslaved workers if nearby host colonies benefit from the reduced raiding activity and host populations are locally structured. By genotyping free-living and enslaved colonies significant isolation-by-distance was observed.

A clonal population structure was detected in aquatic bacteria (Sphingomonadaceae) by multilocus sequencing. Isolates from these subpopulations have been subjected to further testing in high-throughput physiological experiments encompassing a wide variety of different substrates. The results suggest that physiologically distinct lineages arise even within single subpopulations.

In the two theoretical projects several tools have been developed to estimate the parameters of models of population structure, demography and natural selection. Specifically, Jaatha, a new composite-likelihood method that incorporates also the recent divergence of populations and intragenic recombination was designed. Similar to ABC this approach uses summary statistics to compare observed and simulated data and estimate model parameters (including the split time between species). This method was applied to the analysis of wild tomatoes, a clade of young diverging species.

The advances in the second theoretical project include a flexible tool for coalescent simulations (called msms) to analyze selection in populations with complex demography and population structure, and an application of this program to selective sweeps under various modes of dominance.