Empirical Tests of Population Genetics Theory


Estimating population genetic structure among populations connected by dispersal (FST and related methods) is one of the most widely used genetic methodologies today. It is concerning to note that prior to this experiment, the method has never been empirically tested in real populations. In an extensive population manipulation experiment I showed that in populations adhering to the Wright-Fisher model displayed population structure significantly less than predicted by the theory. I inferred that these deviations from theory can be explained if the majority of microsatellite loci are affected by balancing selection at linked genes. This implies that this form of selection is far more pervasive throughout the genome than previously assumed and calls into question the long-established paradigm for rates of neutral molecular evolution, underpinning evolutionary genetics. 


Validating Estimates of Population Structure

Clockwise From Top Left: Drosophila melanogaster in the processes of laying an egg (Photo: Louis Tsai).|  Clare Holleley and Michael Whitehead in the Fly Lab.|  Some of the four thousand replicated Drosophila mating trials.|  Simulated metapopulations that were maintained for 32 generations with controlled population sizes and manipulated rates of gene flow.| Below: Drosophila melanogaster polytene chromosome showing no evidence of large inversions.

In this study, I investigated the precision and accuracy of two single-sample effective population size estimation methods (onesamp and LdNe ) by creating replicated and controlled metapopulations of Drosophila melanogaster that adhered to the Wright-Fisher model of population structure. Contrary to expectations, both onesamp and LdNe underestimated Ne and detected no influence of dispersal on the estimate of Ne. This means that the use of onesamp or LdNe to estimate Ne in internally structured metapopulations can downwardly bias the estimate compared to the true value of Ne for the metapopulation.

Effective Population Size in metapopulations

Induced dispersal (translocation, habitat corridors, reintroductions) is sometimes considered to be a low risk strategy because hybrid breakdown is assumed to occur less frequently than hybrid vigour, however there is very little quantitative data to support this assumption. Conversely, wrongly assuming that hybrid breakdown is widespread and common may result in overly restrictive management strategies that exacerbate the negative effects of fragmentation. Here I showed that in the absence of detectable inbreeding, it is possible for both hybrid vigour and hybrid break down to occur in the F1 generation at a low but equal frequency in significantly structured metapopulations. However I also showed that there is minimal risk hybrids experiencing differential fitness (good or bad) if connectivity is maintained between subpopulations.

Publications from these Investigations

1. Holleley CE, Nichols RA, Whitehead MR, Adamack AT, Gunn MR, Sherwin WB (2013). Testing single-sample estimators of effective population size in genetically structured populations. Conservation Genetics. DOI: 10.1007/s10592-013-0518-3.

2. Dewar RC, Sherwin WB, Thomas E, Holleley CE, Nichols RA (2011). Predictions of single-nucleotide polymorphism differentiation between two populations in terms of mutual information. Molecular Ecology, 20: 3156–3166.

3. Holleley CE, Nichols RA, Whitehead MR, Gunn MR, Gupta J, Sherwin WB (2011). Induced dispersal in wildlife management: experimental evaluation of the risk of hybrid breakdown and the benefit of hybrid vigor in the F1 generation. Conservation Genetics12: 31-40.

Population Connectivity and Hybrid fitness

Due to the magnitude of this enormous Drosophila experiment, Clare Holleley supervised several part-time technicians and Michael Whitehead, a full-time Research Assistant who provided support with Drosophila husbandry, high-throughput DNA extraction and multiplex microsatellite genotyping. In addition to assisting with Clare Holleley’s  Drosophila experiment, Michael Whitehead managed the University of New South Wales’s Molecular Ecology and Evolution Facility (MEEF), a core molecular facility. Michael is now a postdoc working on sexually deceptive orchids and their wasp pollinators.

Part-time Fly Technicians trained and supervised by Clare Holleley: Louis Tsai, Jon Chao, Edwin Ho, Wayne Zhao, Cristy Gelling and Joseph Bilman.

Above: FST lower than expected under the controlled experimental conditions.

Collaborators: Bill Sherwin and Richard Nichols

Full list of all collaborators available here.