Comparisons of different demographic hypotheses in an Approximate Bayesian Computation (ABC) framework supported a population bottleneck in the northwestern French clade and paleo-distribution modelling indicate that the populations of this lineage have experienced more severe environmental fluctuations during the last 21 000 years than those from the Iberian Peninsula.
Multiple tests supported population expansions for each group of populations (boldface); one test supported a population bottleneck (underlined).
Maximum likelihood, mismatch and Bayesian analyses supported a severe historical bottleneck and a relatively recent demographic expansion of the Non-Atlantic clade of S. fimbriatum, whereas size of S. squarrosum populations has probably decreased in the past.
Simulated mutation-drift expectation values for the first temporal sample (before 1920, M = 0.794) were much higher than this critical value in equilibrium under the most realistic evolutionary scenarios (Table 4) supporting a preceding bottleneck.
These results support a response preparation bottleneck and can be explained via a neural activation model of preparation.
This is also reflected in the diversity parameters, which we have obtained using the directed sampling approach, and which do not support a recent population bottleneck or founder effect for the Druze population as a whole.
The external/internal branch length test shows significant support for a bottleneck scenario in lineages I and II.
Conversely, with scattered sampling, we never observe support for a bottleneck, irrespective of M; i.e., Δ l n L is close to zero and θ estimates become very large as M decreases (not shown).
Additional support for a bottleneck in lineage II is found in a significant negative value of Tajima's D for prs and an almost significant negative value for sigB; a negative value of Tajima's D indicates that these loci did not evolve neutrally and is suggestive of a selective sweep or a population bottleneck [ 30, 31].
Furthermore, and in contrast to the PSMC framework (Li and Durbin 2011) that qualitatively documents changes in population size, our method explicitly assesses the statistical support for a bottleneck model over a null model (see Table S3 for the pros and cons of alternative demographic inference methods).
