Tributaries showed a positive correlation coefficient with genetic Epigenetics distance from the TL2 population (Table 3). This observation might indicate the isolation of the TL2 population from other populations for a certain geological time, rather than low occurrence of gene flow between TL2 and other populations due to detouring caused by the Lomami River. TL2 shared no haplotypes with other populations and showed quite different coefficients in the correlation analysis (Table 3). Furthermore, the haplotypes of the D clade were found only in this region (Figure 2). Nevertheless, it contained specific haplotypes of the B1 clade coupled with the west cohort. Future studies will be required to elucidate how the B1 haplotypes are shared between east and west regions (Figure 2). These results might be explained not only by prevention of individual migration by existing riverine networks but also by historical separation of habitats associated with paleoenvironmental changes. The TL2 population might have inhabited another refugium at the LGM between the Congo and Lomami rivers [17,23]. Present-day rivers as barriers to gene flow could not fully explain the genetic structure of bonobo populations confirmed in this study. The geographical pattern of the bonobo genetic structure seems to have formed over hundreds of thousands of years. After bonobos and chimpanzees diverged about 1 Ma [3?], the common ancestor of extant bonobos lived until as recently as 500,000 years ago [1,24]. Even at 500,000 years ago, differentiation of some clades of bonobos occurred long before the LGM (Figure 2). This means that bonobos were affected not only by forest reduction in the LGM but also by climate changes during the Pleistocene, such as the glacial nterglacial pattern. More information on paleoenvironmental 11967625 changes in the Congo Basin during the Pleistocene is required to elucidate the genetic structure of bonobo populations.r (with number of tributaries)nsnsnsnsns ns 0.84 * ns 0.64 ns ns 0.78 0.81 0.88 0.To other five areas (TL2 was removed from calculations) (n = 5)r (with detoured distance )ns0.nsns20.0.0.0.r (with straight distance)**ns**0.r (with number of tributaries)nsns0.nsnsns0.20.0.0.r (with detoured distance )ns0.ns0.20.0.0.ns0.r (with straight distance)To other six areas (n = 6)ns**ns*20.*0.0.0.0.0.*0.0.ns**0.*0.0.*0.nsns0.0.0.0.*Conservation of BonobosIn this study, we classified the bonobo populations in the DRC into three cohorts in different localities (Figure 1). Strong segregation of the cohorts was supported by the observed mtDNA diversity, and they can be regarded as potential evolutionarily significant units in conservation applications [25]. In addition, the geographical distribution of the six clades might reflect differences in evolutionary backgrounds among study populations. To defineWambaSalongaLac TumbaLomakoIyondjiAreaMaleboTLGenetic Structure of BonobosTable 4. Calculations of AIC using GLM for single factor models.FactorAll areas (n = 21) t p 0.000175 0.0000473 0.03571 AIC 216.74 219.51 25.When TL2 was removed (n = 15) t 6.6 (+) 3.1 (+) 3.8 (+) p 0.0000169 0.00905 0.00215 AIC 223.42 29.49 212.Straight distance Detoured 
distance Number of tributaries4.7 (+) 5.2 (+) 2.3 (+)FST was used as a response variable and Gaussian (identity) was used as a family (link function). Signs in parenthesis mean direction to increase FST. doi:10.1371/journal.pone.0059660.tthe species-level diversity of bonobos further, future studies should inhibitor include sample.Tributaries showed a positive correlation coefficient with genetic distance from the TL2 population (Table 3). This observation might indicate the isolation of the TL2 population from other populations for a certain geological time, rather than low occurrence of gene flow between TL2 and other populations due to detouring caused by the Lomami River. TL2 shared no haplotypes with other populations and showed quite different coefficients in the correlation analysis (Table 3). Furthermore, the haplotypes of the D clade were found only in this region (Figure 2). Nevertheless, it contained specific haplotypes of the B1 clade coupled with the west cohort. Future studies will be required to elucidate how the B1 haplotypes are shared between east and west regions (Figure 2). These results might be explained not only by prevention of individual migration by existing riverine networks but also by historical separation of habitats associated with paleoenvironmental changes. The TL2 population might have inhabited another refugium at the LGM between the Congo and Lomami rivers [17,23]. Present-day rivers as barriers to gene 
flow could not fully explain the genetic structure of bonobo populations confirmed in this study. The geographical pattern of the bonobo genetic structure seems to have formed over hundreds of thousands of years. After bonobos and chimpanzees diverged about 1 Ma [3?], the common ancestor of extant bonobos lived until as recently as 500,000 years ago [1,24]. Even at 500,000 years ago, differentiation of some clades of bonobos occurred long before the LGM (Figure 2). This means that bonobos were affected not only by forest reduction in the LGM but also by climate changes during the Pleistocene, such as the glacial nterglacial pattern. More information on paleoenvironmental 11967625 changes in the Congo Basin during the Pleistocene is required to elucidate the genetic structure of bonobo populations.r (with number of tributaries)nsnsnsnsns ns 0.84 * ns 0.64 ns ns 0.78 0.81 0.88 0.To other five areas (TL2 was removed from calculations) (n = 5)r (with detoured distance )ns0.nsns20.0.0.0.r (with straight distance)**ns**0.r (with number of tributaries)nsns0.nsnsns0.20.0.0.r (with detoured distance )ns0.ns0.20.0.0.ns0.r (with straight distance)To other six areas (n = 6)ns**ns*20.*0.0.0.0.0.*0.0.ns**0.*0.0.*0.nsns0.0.0.0.*Conservation of BonobosIn this study, we classified the bonobo populations in the DRC into three cohorts in different localities (Figure 1). Strong segregation of the cohorts was supported by the observed mtDNA diversity, and they can be regarded as potential evolutionarily significant units in conservation applications [25]. In addition, the geographical distribution of the six clades might reflect differences in evolutionary backgrounds among study populations. To defineWambaSalongaLac TumbaLomakoIyondjiAreaMaleboTLGenetic Structure of BonobosTable 4. Calculations of AIC using GLM for single factor models.FactorAll areas (n = 21) t p 0.000175 0.0000473 0.03571 AIC 216.74 219.51 25.When TL2 was removed (n = 15) t 6.6 (+) 3.1 (+) 3.8 (+) p 0.0000169 0.00905 0.00215 AIC 223.42 29.49 212.Straight distance Detoured distance Number of tributaries4.7 (+) 5.2 (+) 2.3 (+)FST was used as a response variable and Gaussian (identity) was used as a family (link function). Signs in parenthesis mean direction to increase FST. doi:10.1371/journal.pone.0059660.tthe species-level diversity of bonobos further, future studies should include sample.
