Extended Data Fig. 8: Summary statistics of allele frequency change between subpopulations. | Nature Ecology & Evolution

Extended Data Fig. 8: Summary statistics of allele frequency change between subpopulations.

From: Experimental evolution of adaptive divergence under varying degrees of gene flow

Extended Data Fig. 8

Data are shown for the local mating (LM) and sympatry (SYM) treatment and are displayed for both ancestral genetic backgrounds (α and β). For each of the 11 populations per treatment the allele frequencies of segregating variants were compared between subpopulations including the following comparisons. Left panel (Top-Bottom): Comparison between the top and bottom ecotypic fractions; comparison between the population pool (prior to ecological selection) with the top (Middle panel: Pool-Top) or bottom ecotypic fraction (Right panel: Pool -Bottom). For each population the proportion of genetic variants with an allele frequency shift of > 0.2 were calculated. The resulting proportion of genetic variants changing in either direction as displayed for each population in Supplementary Fig. 4 are here summarized to show the general trend. Following the colour scheme from Supplementary Fig. 4 boxplots are coloured by the fraction showing an increase in frequency, viz. blue for the top ecotype (diff_Top), red for the bottom ecotype (diff_Bottom) or orange for the pool (diff_Pool). In the α background sympatric populations showed higher similarity between the whole pool sample and the top ecotype fraction (higher proportion of variants increased in top fraction, lower proportion of variants differing between top and the whole pool sample, and higher proportion of variants differing between bottom and whole pool sample). In local mating populations, the pattern was reversed showing higher similarity with the bottom ecotype fraction. In the β background, showed a similar trend, but the difference between treatments was lower, which is consistent with no observed divergence in the phenotypic data. Significance of the difference between groups was tested using a quasibinomial model in a nested generalised lineal model with treatment and fraction as fixed variables. Significant difference between treatments and group are shown with blue and black asterisks respectively. Boxplots as in Extended Data Fig. 1.

Back to article page