Table 2.

Inferred parameters of the DFE for the empirical C. elegans dataset using DFE-alpha

Dataset	Genome portion	SNP set	Population size after first epoch, in 2-epoch mode (N2, relative to N1 = 100)	Duration of epoch after first population size change (t2)	Weighted recent effective population size (Nw, Eyre-Walker and Keightley 2009)	Mean selection coefficient (Es)	Shape parameter (b)
Empirical full strain set	Whole	4-0-fold	1000	299.23	225.06	−40.42	0.17
	Arms		1000	245.70	204.04	−124.88	0.16
	Centers		1000	298.62	224.83	−22.36	0.15
	Whole	4-0-fold, high-moderate	1000	299.23	225.06	−6.80	0.17
	Arms		1000	245.70	204.04	−20.83	0.16
	Centers		1000	298.62	224.83	−2.95	0.15
Empirical swept strain set	Whole	4-0-fold	1000	679.00	359.09	−6.63	0.20
	Arms		1000	628.18	342.59	−197.65	0.14
	Centers		1000	501.51	299.61	−0.72	0.26
	Whole	4-0-fold, high-moderate	1000	679.00	359.09	−1.45	0.20
	Arms		1000	628.18	342.59	−29.71	0.13
	Centers		1000	501.51	299.61	−0.18	0.28
Empirical diverged strain set	Whole	4-0-fold	307	140.07	142.22	−546.09	0.13
	Arms		372	143.18	147.62	−321.05	0.15
	Centers		1000	1814.82	636.79	−16.75	0.14
	Whole	4-0-fold, high-moderate	307	140.07	142.22	−71.53	0.13
	Arms		372	143.18	147.62	−50.33	0.15
	Centers		1000	1814.82	636.79	−2.37	0.13

The mean selection coefficient (Es) is not scaled by effective population size (i.e., not N_es). Mean (absolute) selection coefficients greater than 1 reflect the long tail of a leptokurtic gamma distribution. See Keightley and Eyre-Walker (2007) for explanation and further interpretation. Note that the DFE-alpha approach caps N2 at 1000, reflecting the poor ability to estimate the DFE in these cases due to poor fit during the rescaling.