Skip to main content


Figure 1 | BMC Genetics

Figure 1

From: Automating approximate Bayesian computation by local linear regression

Figure 1

Estimation of bottleneck parameters for European populations of Drosophila melanogaster. The data analyzed are described in [11]. The regression ABC was performed with both tangent [24] and logarithmic transformations [12]. In each panel, the solid line is the approximate posterior distribution obtained using the regression-ABC algorithm and the natural-log transformation, the dotted line is the result of regression-ABC using the transformation from [24], and the dot-dashed line are the rejection sampling results from [11]. The parameters are (a), t r the recovery time from the bottleneck, in units of 4N e generations, (b) d, the duration of the bottleneck in units of 4N e generations, and (c) f, the severity of the bottleneck, which is the ratio of the bottlenecked population size to the pre-bottleneck population size.

Back to article page