Fig 1.
Map of sampled locations in this study.
All localities included in this study are indicated. Terrestrial Ecoregions from the World Wildlife Fund are labeled (Eastern Arc Mountains, South Malawi montane forest–grassland mosaic, Southern Rift montane forest–grassland mosaic). Colors match genomic population clusters. Green = northern Tanzanian populations from the northern Eastern Arc Mountains; Yellow = Udzungwa Mountains population from the southern Eastern Arc Mountains. Black = Livingstone Mountains populations and a coastal Tanzanian locality; Blue = populations in the Southern Highlands; Red = northern Malawian populations; Brown = populations in the Serra Jeci Mountains; White = populations in central Malawi; Purple = populations in southern Malawi and Mozambique. Localities: A. West Usambara Mts, B. East Usambara Mts, C. Nguu Mts, D. Nguru Mts, E. Rubeho, F. Uluguru Mts, G. Malundwe Peak, H. Udzungwa Mts, I. Livingstone Mountains, J. Coastal (Lindi Plateau), K. Southern Highlands, L. Misuku Hills, M. Nyika Plateau, N. Vipya Mts, P. Serra Jeci Mts (Njesi Highlands), Q. Namizimu Forest, R. Zomba Plateau, S. Mt Ribaue, T. Mt Inago, U. Mt Namuli, V. Mt Sokon, W. Mt. Mabu, X. Mt. Mulanje, Y. Cholo Mt, Z. Mt. Chiperone. Map created in ESRI with relief basemap (Copyright:© 2009 ESRI).
Table 1.
Sample sizes for all groups (number of individuals).
Fig 2.
Phylogeographic clades of mitochondrial and genomic loci.
A. BEAST2 Bayesian Inference phylogenetic tree of mitochondrial ND2 dataset. Posterior probabilities shown at nodes with branch lengths proportional to nucleotide divergence (asterixis indicated 1.0 posterior values). Names match the locations indicated by letters in Fig 1 and can be referenced from the figure legend. Colors match localities in Fig 1 (based on genomic results) except in the Udzungwa mountain block (yellow in map figure) which are shown in orange and yellow to represent the two unconnected mitochondrial clades (northern Udzungwa in orange and southern Udzungwa in yellow). Mitochondrial and nuclear haplotype networks (ND2, 16S, POMC, and c-Myc) are in supplementary materials. Arcs connecting clades indicate relationships supported by nuclear loci (c-Myc and POMC) that are in discordance with the ND2 mitochondrial tree (S1 Fig). B. Distance tree (proportion of loci that are different) based on genomic dataset. All nodes show 100 bootstrap support, with asterisks displayed on major nodes to denote this. C. SNAPP tree of SNPs. The figure shows the posterior distribution of species tree topologies based on the multi-species coalescent approach with the most likely tree shown in blue and alternate topologies shown in green and red. All major nodes have 1.0 posterior probability, indicated by asterisks. D. Hydrological drainage basins (hydrobasins) and major phylogenetic clades are shown. The three main phylogenetic clusters in the genomic dataset are identified as Clades 1–3. Major hydrobasins were obtained from www.hydrosheds.org, downloaded on February 2011 © HydroSHEDS. Inset: A closer map of the region in south-central Tanzania and norther Malawi is shown encompassing populations that have unexpected mitochondrial breaks given the close spatial proximity. The two mitochondrial haplotypes in the Udzungwa Mountains (yellow and orange) are in two different minor hydrobasins. Likewise, the two central red populations on the north and south slopes of the Nyika Plateau cluster with the other sampling locations of each shared minor hydrobasin. Arrows pointing to the spatial location of the nearest mitochondrial connection for these four localities are shown.
Fig 3.
A. Principal component analysis (PCA) on SNP dataset. Each point represents the SNP profile of an individual. The X axis represents the variance explained by PC1 (29%), and Y axis represents the variance explained by PC2 (16%). Ellipses include 95% of the data for each population. B. DAPC analysis with PCA eigenvalues and DA eigenvalues inset. C. Treemix estimate of unrooted phylogenetic tree with migration. Sizable migration was inferred across many branches of the phylogenetic tree (orange and red arrows). Each population is represented by a colored circle. Coloration of populations matches STRUCTURE plots (the three black localities in Fig 1 are shown as yellow outlined with black as in Fig 2D). D. Genetic structure plot from STRUCTURE (K = 7). Each vertical bar corresponds to one individual. E. Photos of individuals from each population except Southern Highlands. Colors match localities in Fig 1, except the coastal Tanzania and Southern Highlands group (black in Fig 1) are colored as yellow with black outline due to STRUCTURE clustering with Udzungwa Mountain samples (yellow).