TABLE 4

Relative maximum-likelihood (ML) and moment (MT) estimates of 1/(2Ne) and m for quasi-infinite populations

ParametersRelative ML estimates of 1/(2Ne)Relative ML estimates of mRelative MT estimates of 1/(2Ne)Relative MT estimates of m
NemMeanSD2.5%97.5%MeanSD2.5%97.5%MeanSD2.5%97.5%MeanSD2.5%97.5%
Island model
100.011.0410.3140.4931.7091.3532.1990.0057.2710.9740.4180.3041.9062.0284.680-6.11711.70
0.10.9160.3310.3931.6990.9560.6610.0022.4210.9480.4080.2891.8141.0720.806-0.2392.880
0.50.8730.3460.3241.5720.9020.4000.2541.9990.8880.4400.2091.9150.9440.3630.3041.703
1000.011.0840.2450.6721.5911.0180.6680.0112.5330.9720.3310.3801.7172.1221.2440.0284.786
0.11.0120.2410.5551.5290.8600.3540.2291.5810.8370.2700.3491.3901.4530.3410.8622.190
0.51.0440.2570.6041.6190.9030.2440.4611.4650.3950.2630.0000.9491.0450.1860.7381.460
Circular stepping-stone model
100.011.2620.3200.5951.8971.2911.8050.0056.3250.9260.4540.2082.0202.1064.404-5.10111.80
0.11.0700.3260.5111.8381.1310.7720.0012.9470.8870.4650.1211.9281.3331.189-0.5793.874
0.50.9300.3380.3481.7111.0010.4770.2141.9990.9000.4790.1382.0400.9570.4190.2211.787
1000.011.0780.2280.6901.5441.1310.7890.0082.8870.9890.3610.3771.7682.5281.433-0.0655.587
0.11.0300.2350.6041.5420.9950.4210.2321.8940.8130.3330.1961.5431.7020.4370.8972.623
0.50.9770.2510.5111.5040.9940.3370.4521.9950.3760.2860.0000.9361.0420.2190.6951.623
  • The simulated metapopulation contains 21 populations of equal Ne in an island model or a circular stepping-stone model. The focal populations have various effective sizes (first column) and immigration rates (second column). Twenty loci, each having an equal number of alleles with initial frequencies drawn from a uniform distribution, are used in the estimation. Two samples of equal size separated by two generations are taken from the focal and each of four source populations (the island model) or two neighboring populations (the circular stepping-stone model). For populations with Ne = 10 and 100, the sizes of the samples are 50 and 100 diploid individuals, respectively, and the numbers of alleles per locus are two and eight, respectively. In each case, 1000 replicates are run and the estimation is summarized by the mean, standard deviation, and 2.5 and 97.5 percentiles of the estimates of 1/(2Ne) and m relative to their true values.