Table 3. We infer more nearly neutral (|s| < 10⁻⁵) and fewer strongly deleterious (|s| ≥ 10⁻²) new mutations than previous studies.

Data set	μ, LNS/LS	Best fit DFE	0 ≤ |s| < 10−5	10−5 ≤ |s| < 10−4	10−4 ≤ |s| < 10−3	10−3 ≤ |s| < 10−2	10−2 ≤ |s|
Boyko et al. (2008) (AAa)	1.8 × 10−8, 2.5b	Gamma	0.183	0.096	0.147	0.219	0.355
1000 Genomes	1.8 × 10−8, 2.5b	Gamma	0.217 (0.212–0.223)	0.112 (0.111–0.113)	0.169 (0.165–0.172)	0.243 (0.235–0.249)	0.259 (0.252–0.266)
ESP		Gamma	0.229 (0.223–0.234)	0.105 (0.104–0.106)	0.152 (0.150–0.155)	0.216 (0.212–0.221)	0.298 (0.294–0.302)
LuCamp		Discrete	0.278 (0.221–0.303)	0.027 (0.001–0.110)	0.211 (0.167–0.234)	0.352 (0.330–0.373)	0.132 (0.124–0.142)
1000 Genomes	1.5 × 10−8, 2.31	Gamma	0.237 (0.231–0.243)	0.127 (0.125–0.128)	0.192 (0.188–0.197)	0.266 (0.259–0.272)	0.178 (0.171–0.186)
ESP		Neu+gamma	0.263 (0.250–0.277)	0.104 (0.091–0.114)	0.167 (0.160–0.173)	0.249 (0.241–0.259)	0.217 (0.211–0.221)
LuCamp		Neu+gamma	0.242 (0.223–0.260)	0.091 (0.072–0.107)	0.194 (0.183–0.203)	0.332 (0.313–0.352)	0.141 (0.129–0.152)

Results in comparison with data from Boyko et al. (2008). C.I.’s were constructed by Poisson resampling the nonsynonymous SFS and fitting a DFE 200 times. See Table S5 in File S1 for additional information. Neu, neutral.

^aAfrican-American.

^bThe results presented with the assumptions L_NS/L_S = 2.5 and μ = 1.8 × 10⁻⁸ match the mutation rate assumptions of Boyko et al. (2008).