261
The phenotypic differentiation (Q ST ) ranged from near cero for SPL and NLP to
0.36 for NPI, with an average over traits of 0.14. Molecular differentiation among
origins was low but significant (F ST = 0.069, p F ST whereas for the remaining five
traits (NLP, PIL, SPL, LEF, LEX/LEA) Q ST – F ST < 0. Cases where Q ST was higher
than the upper limit of F ST confidence interval were considered as suggestive of
directional selection while the cases where Q ST was lower than the lower F ST confidence interval limit are indicative of stabilizing selection. As discussed in Bessega
et al. (2009), the univariate tests suggest heterogeneous selection over phenotypic
traits with different optima in different populations. This result is in agreement with
several authors that described the occurrence of ecotypes adapted to particular areas
in several species of the genus Prosopis (Morello et al. 1971; Burghardt et al. 2004;
Verga et al. 2009, 2014). The heterogeneous selection over phenotypic traits has
important implications for gene resource management, suggesting that population
source and traits with adaptive significance should be considered in breeding programs as a crucial point prior to any selecting decision.
9.4 Estimation of Mating System and Pollen
Dispersal Parameters
Mating system and pollen dispersal parameters can be estimated indirectly in different populations based on molecular markers (Ritland 1990; Austerlitz and Smouse
2001; Smouse et al. 2001). An efficient method based on the mixed multilocus
mating model was developed by Ritland (2002) and is implemented in the MLTR
program (Ritland 2002). The method requires the grouping of the population
samples into families (seeds collected from the same tree) and the multilocus
genotyping of corresponding individuals. The comparison of the allelic frequencies
in the pollen cloud that attends each mother plant allows, among other parameters,
to establish the outcrossing rate (t), the biparental inbreeding (tm – ts), the correlated
outcrossing rate (rt), the correlated paternity (rp), the pollen structure parameter
(Φ ft ), and the effective number of pollen donors siring each mother plant (Nep).
On the basis of genetic data, it is also possible to indirectly assess the distance (in
meters) that pollen is dispersed in populations. From the genotypes of the mother
plants and their offspring (seeds collected individually from each tree), it is possible
to estimate again the composition of the pollen cloud and indirectly estimate the
pollen dispersion parameters using the computer program POLDISP 1.0 (RobledoArnuncio et al. 2007), applying two different algorithms: TwoGener (Smouse et al.
2001; Austerlitz and Smouse 2002) and Kindist (Robledo-Arnuncio et al. 2006).
Both methods are based on the expected relationship between the physical distance
between the sampled mother plants and the interclass genetic correlation (Φ FT )
between pairs of individuals (seeds) for the pollen set sampled in each mother.
The coefficient Φ FT can be considered an estimator of kinship, and it is expected that
the kinship between the descendants of different mother plants is inversely
9 Genetic Variation Patterns of “Algarrobos” from the “Great American Chaco…
The phenotypic differentiation (Q ST ) ranged from near cero for SPL and NLP to
0.36 for NPI, with an average over traits of 0.14. Molecular differentiation among
origins was low but significant (F ST = 0.069, p F ST whereas for the remaining five
traits (NLP, PIL, SPL, LEF, LEX/LEA) Q ST – F ST < 0. Cases where Q ST was higher
than the upper limit of F ST confidence interval were considered as suggestive of
directional selection while the cases where Q ST was lower than the lower F ST confidence interval limit are indicative of stabilizing selection. As discussed in Bessega
et al. (2009), the univariate tests suggest heterogeneous selection over phenotypic
traits with different optima in different populations. This result is in agreement with
several authors that described the occurrence of ecotypes adapted to particular areas
in several species of the genus Prosopis (Morello et al. 1971; Burghardt et al. 2004;
Verga et al. 2009, 2014). The heterogeneous selection over phenotypic traits has
important implications for gene resource management, suggesting that population
source and traits with adaptive significance should be considered in breeding programs as a crucial point prior to any selecting decision.
9.4 Estimation of Mating System and Pollen
Dispersal Parameters
Mating system and pollen dispersal parameters can be estimated indirectly in different populations based on molecular markers (Ritland 1990; Austerlitz and Smouse
2001; Smouse et al. 2001). An efficient method based on the mixed multilocus
mating model was developed by Ritland (2002) and is implemented in the MLTR
program (Ritland 2002). The method requires the grouping of the population
samples into families (seeds collected from the same tree) and the multilocus
genotyping of corresponding individuals. The comparison of the allelic frequencies
in the pollen cloud that attends each mother plant allows, among other parameters,
to establish the outcrossing rate (t), the biparental inbreeding (tm – ts), the correlated
outcrossing rate (rt), the correlated paternity (rp), the pollen structure parameter
(Φ ft ), and the effective number of pollen donors siring each mother plant (Nep).
On the basis of genetic data, it is also possible to indirectly assess the distance (in
meters) that pollen is dispersed in populations. From the genotypes of the mother
plants and their offspring (seeds collected individually from each tree), it is possible
to estimate again the composition of the pollen cloud and indirectly estimate the
pollen dispersion parameters using the computer program POLDISP 1.0 (RobledoArnuncio et al. 2007), applying two different algorithms: TwoGener (Smouse et al.
2001; Austerlitz and Smouse 2002) and Kindist (Robledo-Arnuncio et al. 2006).
Both methods are based on the expected relationship between the physical distance
between the sampled mother plants and the interclass genetic correlation (Φ FT )
between pairs of individuals (seeds) for the pollen set sampled in each mother.
The coefficient Φ FT can be considered an estimator of kinship, and it is expected that
the kinship between the descendants of different mother plants is inversely
9 Genetic Variation Patterns of “Algarrobos” from the “Great American Chaco…