Discriminant
Analysis of Principal Components (DAPC)
We ran a DAPC analysis to
optimize the variation between populations while reducing within
population variation. When grouping individuals by continent, we were
able to detect the variation among samples from different continents
(Figure 6A). Both the African and Asian groups were separated from the
rest of the samples, indicating greater variation among those lice. Even
though European, South + Central American and North American samples
formed different clusters, the variation among those was small as
compared to the African and Asian groups (Figure 6A). We also ran the
DAPC analysis without adding any prior information about the sampling
localities and tested putative genetic populations K=1 through K=50. We
retained K=5 (Figure 6B) as the optimum K after using the find.clusters
parameter and calculating the BIC scores for each K tested. This K=5
genetic clustering agrees with the results obtained from both the
fastSTRUCTURE and PCA analyses, and the distribution of individuals
(Figure 6B) was also similar to that of the PCA and fast STRUCTURE
results. In addition, like the PCA results, discriminant component 1
separated the sub-Saharan African samples from the rest. North African
(Algeria and Egypt) samples were separated from this main sub-Saharan
African cluster