Population Genetic Structure, Dispersal & Gene Flow in the German Cockroach, Blattella Germanica: A Structural Pest in the Urban Environment

Abstract

The goal of the current study was to explore German cockroach population differentiation and dispersal in urban residences at several spatial scales. Little differentiation was expected within buildings due to frequent active dispersal of the German cockroach. However, differentiation was expected between spatially separated buildings because the German cockroach cannot survive beyond human structures and dispersal must be human-mediated. Differentiation was expected to increase with increasing spatial distance as human-mediated dispersal becomes successively less frequent. Cockroaches were collected in a geographical hierarchy, with samples at the scale of single apartments, apartment buildings, and apartment complexes within Raleigh, North Carolina in addition to 16 other cities in the United States and six Eurasian cities. All individuals were genotyped at ten microsatellite loci. Microsatellite diversity was generally high across our samples, with average alleles per locus as high as 9.11, and observed heterozygosity in excess of 0.6 in most populations. However, global differentiation was lower than expected, with a global FST of 0.046 across Raleigh, 0.099 across the U.S. with Raleigh excluded, and 0.158 across global samples with Raleigh excluded. No departure from panmixia was found between German cockroach samples within apartments by pairwise G-tests (P > 0.05). G-tests did detect highly significant departures from panmixia between all but one pair of apartments within Raleigh (all P < 0.001). These tests indicate that each separate apartment represents an individual population. Global FST values suggested that dispersal and gene flow occurred more often within buildings than between them and more within apartment complexes than between them. Similarly, significant Mantel tests of isolation by distance within Raleigh (P < 0.001), across the U.S. (P = 0.029), and across global samples (P < 0.005) indicated a significant positive correlation between genetic distance and geographic distance. These tests suggested that German cockroach dispersal and gene flow occurred less frequently at successively larger spatial scales. However, phylogenetic trees constructed from genetic distances for each spatial scale were unable to cluster apartments according to geographic proximity, except for apartments within a single building. The individual-based clustering program Structure was also unable to detect any informative underlying structure beyond a single building when no a priori population information was used. A similar lack of geographic pattern was shown in B. germanica mitochondrial 16S diversity. Only 10 unique 16S haplotypes were found for 66 individuals across 40 global populations, and haplotypes varied greatly in their frequency and geographic ranges. Beyond the scale of a single building, gene flow by passive dispersal seems to have formed no spatial clustering of genetic similarity, even at the global scale. This may be due to non-equilibrium conditions and non-linear evolution caused by dramatic declines in population size due to insecticides or large fluctuations in temperature, humidity, and food availability. These population bottlenecks are known to rapidly change allele frequencies and affect various evolutionary estimates, including genetic distance. The challenge for future studies will be to find markers conserved enough to be robust to these non-equilibrium conditions but variable enough for fine scale analysis.