356
In summary, genetic diversity of C. balansae populations in the Yungas of NWA
was moderate, with a similar general trend displayed by both markers, making the
results comparable. Levels of genetic diversity observed in C. balansae might be
associated with the species’ marginal distribution area and the latitudinal position of
the studied populations on the southernmost edge of a major ecosystem as the
Neotropical Cloud Forests. The intense exploitation, to which populations have
been subjected due to its hardwood qualities, easy accessibility, and species’ spatial
distribution pattern, also accounts for its levels of genetic diversity.
Population genetic differentiation, by SSR and AFLP data, was estimated using
θp and Fst statistics. Historical gene flow (Nm) was then indirectly calculated,
according to Crow and Aoki (1984). Bayesian clustering approach was implemented
for both markers, to resolve the optimal number of genetic clusters (K) and population assignment across the identified clusters. The no-admixture model was used in
order to detect subtle structure (Structure 2.3.3 software, Pritchard et al. 2000;
Falush et al. 2007). Low but significant genetic differentiation among populations
(SSRs: θp = 0.049, CI95: [0.025; 0.077]; AFLPs: Fst = 0.041, P ≤ 0.001) was
observed. Additionally, considerable historical gene flow was detected among
C. balansae populations, with Nm values of 3.71 and 4.47, for SSRs and AFLPs,
respectively. Regarding population structure, four genetic clusters (K = 4, according
to Evanno et al. 2005 method) homogeneously distributed were distinguished using
Bayesian methods. No pattern of clustering between the populations was observed,
and there were no populations assigned exclusively to a given genetic cluster
(Fig. 13.3b).
A cluster analysis of individuals based on similarity matrices was also implemented using UPGMA method. Mantel tests were applied and cophenetic correlation coefficients (r) were computed; significance testing was achieved with 1000
random permutations using NTSYS pc.2.0 software. The 107 individuals analyzed
could be discriminated using the AFLPs markers, whereas the SSR set did not allow
to distinctively fingerprint 7 pairs of individuals (data not shown). No grouping by
population nor by geographic location was observed in either case, being consistent
with genetic cluster distribution. Moderate to high values were obtained for the
cophenetic correlation coefficients (r = 0.676 and r = 0.884 for SSRs and AFLPs,
respectively). Very low genetic distances between populations were observed, being
the largest 0.047 and 0.070 for SSRs and AFLPs (r = 0.780 and r = 0.740), respectively. Finally, a very low correlation between geographic and genetic distances was
observed for both molecular markers (SSRs, r = 0.197, P = 0.01; AFLPs, r = 0.056,
P = 0.01), and no spatial pattern of distribution was detected. The weak grouping of
individuals was further confirmed by cluster analysis and the extremely low pairwise genetic distances recorded.
Taking into account the short geographic distance among C. balansae populations in the Yungas (~250 km between the northernmost and the southernmost ones),
it is likely that they are in a potential reproductive contact which is being reflected
in a weak population structure and clustering of individuals based on similarity
indices. In addition, the high levels of historical gene flow detected among populations would suggest that the 8 populations of C. balansae studied behave practically
as a homogeneous genetic unit in the piedmont of the Yungas.
N. Zelener et al.
In summary, genetic diversity of C. balansae populations in the Yungas of NWA
was moderate, with a similar general trend displayed by both markers, making the
results comparable. Levels of genetic diversity observed in C. balansae might be
associated with the species’ marginal distribution area and the latitudinal position of
the studied populations on the southernmost edge of a major ecosystem as the
Neotropical Cloud Forests. The intense exploitation, to which populations have
been subjected due to its hardwood qualities, easy accessibility, and species’ spatial
distribution pattern, also accounts for its levels of genetic diversity.
Population genetic differentiation, by SSR and AFLP data, was estimated using
θp and Fst statistics. Historical gene flow (Nm) was then indirectly calculated,
according to Crow and Aoki (1984). Bayesian clustering approach was implemented
for both markers, to resolve the optimal number of genetic clusters (K) and population assignment across the identified clusters. The no-admixture model was used in
order to detect subtle structure (Structure 2.3.3 software, Pritchard et al. 2000;
Falush et al. 2007). Low but significant genetic differentiation among populations
(SSRs: θp = 0.049, CI95: [0.025; 0.077]; AFLPs: Fst = 0.041, P ≤ 0.001) was
observed. Additionally, considerable historical gene flow was detected among
C. balansae populations, with Nm values of 3.71 and 4.47, for SSRs and AFLPs,
respectively. Regarding population structure, four genetic clusters (K = 4, according
to Evanno et al. 2005 method) homogeneously distributed were distinguished using
Bayesian methods. No pattern of clustering between the populations was observed,
and there were no populations assigned exclusively to a given genetic cluster
(Fig. 13.3b).
A cluster analysis of individuals based on similarity matrices was also implemented using UPGMA method. Mantel tests were applied and cophenetic correlation coefficients (r) were computed; significance testing was achieved with 1000
random permutations using NTSYS pc.2.0 software. The 107 individuals analyzed
could be discriminated using the AFLPs markers, whereas the SSR set did not allow
to distinctively fingerprint 7 pairs of individuals (data not shown). No grouping by
population nor by geographic location was observed in either case, being consistent
with genetic cluster distribution. Moderate to high values were obtained for the
cophenetic correlation coefficients (r = 0.676 and r = 0.884 for SSRs and AFLPs,
respectively). Very low genetic distances between populations were observed, being
the largest 0.047 and 0.070 for SSRs and AFLPs (r = 0.780 and r = 0.740), respectively. Finally, a very low correlation between geographic and genetic distances was
observed for both molecular markers (SSRs, r = 0.197, P = 0.01; AFLPs, r = 0.056,
P = 0.01), and no spatial pattern of distribution was detected. The weak grouping of
individuals was further confirmed by cluster analysis and the extremely low pairwise genetic distances recorded.
Taking into account the short geographic distance among C. balansae populations in the Yungas (~250 km between the northernmost and the southernmost ones),
it is likely that they are in a potential reproductive contact which is being reflected
in a weak population structure and clustering of individuals based on similarity
indices. In addition, the high levels of historical gene flow detected among populations would suggest that the 8 populations of C. balansae studied behave practically
as a homogeneous genetic unit in the piedmont of the Yungas.
N. Zelener et al.