How Ebola impacts genetics of Western lowland gorilla populations

Abstract

Background: Emerging infectious diseases in wildlife are major threats for both human health and biodiversity conservation. Infectious diseases can have serious consequences for the genetic diversity of populations, which could enhance the species' extinction probability. The Ebola epizootic in western and central Africa induced more than 90% mortality in Western lowland gorilla population. Although mortality rates are very high, the impacts of Ebola on genetic diversity of Western lowland gorilla have never been assessed.

Methodology/principal findings: We carried out long term studies of three populations of Western lowland gorilla in the Republic of the Congo (Odzala-Kokoua National Park, Lossi gorilla sanctuary both affected by Ebola and Lossi's periphery not affected). Using 17 microsatellite loci, we compared genetic diversity and structure of the populations and estimate their effective size before and after Ebola outbreaks. Despite the effective size decline in both populations, we did not detect loss in genetic diversity after the epizootic. We revealed temporal changes in allele frequencies in the smallest population.

Conclusions/significance: Immigration and short time elapsed since outbreaks could explain the conservation of genetic diversity after the demographic crash. Temporal changes in allele frequencies could not be explained by genetic drift or random sampling. Immigration from genetically differentiated populations and a non random mortality induced by Ebola, i.e., selective pressure and cost of sociality, are alternative hypotheses. Understanding the influence of Ebola on gorilla genetic dynamics is of paramount importance for human health, primate evolution and conservation biology.

Figure 1. Map of the studied populations.

The locations of the samples collected before and after Ebola outbreaks are showed.

Figure 2. Comparison of genetic diversity indexes between post epidemic and simulated bottlenecked populations.

The comparison of number of alleles, expected and observed heterozygosities between post-epidemic population (black circles) and bottlenecked populations simulated for long-lived overlapping generation species with initial population size (N₀) 5,000 (black squares), 2,500 (open squares), 1,000 (black triangles), 500 (open triangles) are presented for Lossi (A, B, C) and Lokoue (D, E, F). Results shown are from simulation with female biased mortality during bottleneck (sex ratio after bottleneck is 1∶2). Allele frequencies of initial population are the ones of pre-epidemic samples of Lossi and Lokoué. Years: 1: just after the first bottleneck (90% of mortality). 2: the population size is constant i.e. equal to year 1. 3: just after the second bottleneck (90% of mortality). 4: the population size is constant i.e. equal to year 3. ** indicates significant Wilcoxon's signed rank test both between observed values and simulated values and between years after the second bottleneck and years after the first bottleneck. Vertical bars represent standard error.

Figure 3. Distribution of allele frequencies of the loci showing significant temporal and spatial changes in Lossi.

The pre-epidemic (black bars), post-epidemic (grey bars) and periphery (hatched bars) allele frequencies distribution are presented for the two loci with significant Waples' neutrality tests when comparing pre and post epidemic samples (A and B) and for the two loci with significant homogeneity tests when comparing pre, post epidemic and periphery samples (C and D). n: number of individuals typed for the locus.