Data set | DFE | Parameter MLEs | Log-likelihood | AIC | |ΔAIC|^{a} |
---|---|---|---|---|---|

LuCamp | Gamma | α = 0.215, β = 562.1 | −3334.7 | 6673.4 | 13 |

Neu+gamma | p_{neu} = 0.164, α = 0.338, β = 367.7 | −3327.2 | 6660.4 | 0 | |

Neu+exp+let | p_{neu} = 0.304, p_{exp} = 0.613, λ = 66.56 | −3337.8 | 6681.6 | 21.2 | |

Discrete^{b} | p_{1} = 0.309, p_{2} = 0.024, p_{3} = 0.247, p_{4} = 0.372 | −3334.1 | 6676.2 | 15.8 | |

1kG EUR | Gamma | α = 0.186, β = 875.0 | −1450.5 | 2905.0 | 0 |

Neu+gamma | p_{neu} = 0.031, α = 0.199, β = 820.6 | −1450.8 | 2907.6 | 2.6 | |

Neu+exp+let | p_{neu} = 0.312, p_{exp} = 0.509, λ = 41.48 | −1472.0 | 2950.0 | 45 | |

Discrete^{b} | p_{1} = 0.286, p_{2} = 0.099, p_{3} = 0.222, p_{4} = 0.305 | −1453.4 | 2914.8 | 9.8 | |

ESP EUR | Gamma | α = 0.169, β = 1327.4 | −3012.6 | 6029.2 | 2.6 |

Neu+gamma | p_{neu} = 0.092, α = 0.207, β = 1082.3 | −3010.3 | 6026.6 | 0 | |

Neu+exp+let | p_{neu} = 0.341, p_{exp} = 0.504, λ = 63.90 | −3071.6 | 6149.2 | 122.6 | |

Discrete^{b} | p_{1} = 0.334, p_{2} = 0.041, p_{3} = 0.201, p_{4} = 0.306 | −3029.5 | 6067.0 | 40.4 |

These results are reported assuming

*L*_{NS}/*L*_{S}= 2.31 and*μ*= 1.5×10^{−8}. See Table S4 in File S1 for additional information. The shape and scale parameters of the gamma distribution are denoted with α and β, respectively, and the rate parameter of the exponential distribution is denoted with λ. Neu, neutral; exp, exponential; let, lethal; 1kG, 1000 Genomes.↵a Change in AIC relative to the model with the lowest AIC.

↵b In terms of |

*s*|, these parameters correspond to the ranges of |*s*| corresponding to: [0, 10^{−5}), [10^{−5}, 10^{−4}), [10^{−4}, 10^{−3}), and [10^{−3}, 10^{−2}), respectively. The mass in the [10^{−2}, 1] range is the complement of the total mass of the four aforementioned categories.