Molecular tools, optimal sampling strategies and biogeographical investigation towards the study of adaptive gene flow in forest trees

Abstract

My thesis aimed at providing a conceptual and methodological framework for the study of adaptive gene flow in forest trees. The main elements needed to perform an experiment for estimating adaptive gene flow are: an adequate set of molecular markers, a suitable combination of study species × biogeographical zone upon which designing the experiment, and an optimal sampling strategy for identifying relevant areas where it is most likely to detect adaptive gene flow, and for correctly describing gene flow patterns. In my thesis, I investigated these issues endeavoring to make a significant contribution to each of them, focusing on Apennine populations of silver fir (Abies alba Mill.), a European conifer species with high ecological and economical value.

I investigated first the effect of sampling strategy on the estimation of pollen dispersal patterns (i.e. pollen immigration rate and parameters of the dispersal kernel) in paternity studies in order to set up optimal sampling strategies. After conducting a thorough review of published paternity studies, I evaluated whether sampling efforts commonly used in published paternity studies are adequate to estimate pollen dispersal parameters, using a simulation approach. I found that low sampling efforts can result in highly biased and imprecise estimates of the parameters characterizing pollen dispersal patterns.

Second, in order to have an adequate set of molecular markers for performing biogeographical and gene flow studies on silver fir, I developed a new set of 16 polymorphic microsatellites from transcriptome sequencing. These markers were assembled in 2 8-plexes with high amplification success and clear band pattern. A third multiplex of genomic microsatellites was assembled choosing the best available markers and including 2 newly developed genomic microsatellites. In a parallel project a set of 763 potentially adaptive SNPs was also developed (using transcriptome data and candidate genes linked to metabolism, growth and responses to stress) and tested on samples from my biogeographical experiment.

Finally, I focused on the selection of adequate geographical areas to study adaptive gene flow by analyzing the biogeographical structure of silver fir populations along the Apennine range. These populations have high evolutionary and conservation value because they are at the rear-edge of the species distribution and they were shown to include putative glacial refugia, thus representing hotspots of genetic diversity. Furthermore, they grow in a wide variety of environmental conditions, since the Apennines are a 1000 km mountain chain with large environmental heterogeneity. Individuals from 16 populations sampled across the whole Apennine range were genotyped using a set of 17 microsatellites (15 chosen from the ones developed in this thesis work). By using multiple clustering methods, I found clearly separated genetic clusters, probably reflecting the biogeographical history of the species. These results allowed me to identify: i) genetically homogeneous clusters where it is possible to investigate clinal genetic variation along environmental gradients, and ii) geographically circumscribed areas with abrupt genetic discontinuities. These areas of genetic discontinuity maximize the probability of observing adaptive gene flow, if any, and thus they are the ideal candidate sites for performing an experiment focused on the quantification of adaptive gene flow. This experiment can be done by intensifying the sampling grid in order to study the genetic structure at finer scale. Once described the fine-scale component of boundaries among genetic clusters, gene flow between the closest divergent populations can be studied using paternity and parentage analysis, taking advantage of the sampling guidelines and the molecular markers developed in my this thesis.