Does GST underestimate genetic differentiation from marker data?

The widely applied genetic differentiation statistics F_ST and G_ST have recently been criticized for underestimating differentiation when applied to highly polymorphic markers such as microsatellites. New statistics claimed to be unaffected by marker polymorphisms have been proposed and advocated to replace the traditional F_ST and G_ST. This study shows that G_ST gives accurate estimates and underestimates of differentiation when demographic factors are more and less important than mutations, respectively. In the former case, all markers, regardless of diversity (H_S), have the same G_ST value in expectation and thus give replicated estimates of differentiation. In the latter case, markers of higher H_S have lower G_ST values, resulting in a negative, roughly linear correlation between G_ST and H_S across loci. I propose that the correlation coefficient between G_ST and H_S across loci, r_GH, can be used to distinguish the two cases and to detect mutational effects on G_ST. A highly negative and significant r_GH, when coupled with highly variable G_ST values among loci, would reveal that marker G_ST values are affected substantially by mutations and marker diversity, underestimate population differentiation, and are not comparable among studies, species and markers. Simulated and empirical data sets are used to check the power and statistical behaviour, and to demonstrate the usefulness of the correlation analysis.     

Confidently identifying the correct K value using the ΔK method: When does K = 2?

Populations delineated based on genetic data are commonly used for wildlife conservation and management. Many studies use the program structure combined with the ΔK method to identify the most probable number of populations (K). We recently found K = 2 was identified more often when studies used ΔK compared to studies that did not. We suggested two reasons for this: hierarchical population structure leads to underestimation, or the ΔK method does not evaluate K = 1 causing an overestimation. The present contribution aims to develop a better understanding of the limits of the method using one, two and three population simulations across migration scenarios. From these simulations we identified the “best K” using model likelihood and ΔK. Our findings show that mean probability plots and ΔK are unable to resolve the correct number of populations once migration rate exceeds 0.005. We also found a strong bias towards selecting K = 2 using the ΔK method. We used these data to identify the range of values where the ΔK statistic identifies a value of K that is not well supported. Finally, using the simulations and a review of empirical data, we found that the magnitude of ΔK corresponds to the level of divergence between populations. Based on our findings, we suggest researchers should use the ΔK method cautiously; they need to report all relevant data, including the magnitude of ΔK, and an estimate of connectivity for the research community to assess whether meaningful genetic structure exists within the context of management and conservation.     

When can noninvasive samples provide sufficient information in conservation genetics studies?

Noninvasive sampling, of faeces and hair for example, has enabled many genetic studies of wildlife populations. However, two prevailing problems common to these studies are small sample sizes and high genotyping errors. The first problem stems from the difficulty in collecting noninvasive samples, particularly from populations of rare or elusive species, and the second is caused by the low quantity and quality of DNA extracted from a noninvasive sample. A common question is therefore whether noninvasive sampling provides sufficient information for the analyses commonly conducted in conservation genetics studies. Here, we conducted a simulation study to investigate the effect of small sample sizes and genotyping errors on the precision and accuracy of the most commonly estimated genetic parameters. Our results indicate that small sample sizes cause little bias in measures of expected heterozygosity, pairwise F_ST and population structure, but a large downward bias in estimates of allelic diversity. Allelic dropouts and false alleles had a much smaller effect than missing data, which effectively reduces sample size further. Overall, reasonable estimates of genetic variation and population subdivision are obtainable from noninvasive samples as long as error rates are kept below a frequency of 0.2. Similarly, unbiased estimates of population clustering can be made with genotyping error rates below 0.5 when the populations are highly differentiated. These results provide a useful guide for researchers faced with studying the conservation genetics of small, endangered populations from noninvasive samples.