Skip to main content
. 2015 Jan 27;64(3):472–491. doi: 10.1093/sysbio/syv006

Table 4.

A detailed look at performance on data set DS1 with and without MCMCMC using varying number of runs

Ch Runs Peak Conv Lim Iterations Run Time MaxErr RMSD
1 2 0 0 0 2.42E+06 4.84E+06 0.266 0.106
1 3 6 3 0 2.06E+07 6.18E+07 0.108 0.045
1 4 8 3 1 3.39E+07 1.36E+08 0.081 0.036
1 5 9 7 3 5.88E+07 2.94E+08 0.052 0.024
1 6 9 6 3 6.64E+07 3.98E+08 0.053 0.024
1 7 9 8 4 7.52E+07 5.26E+08 0.041 0.018
1 8 9 8 6 8.26E+07 6.61E+08 0.037 0.016
4 2 8 5 0 8.50E+05 6.80E+06 0.082 0.038
4 3 10 6 0 4.05E+06 4.86E+07 0.038 0.018
4 4 10 6 0 4.07E+06 6.51E+07 0.03 0.015
4 5 10 8 0 6.52E+06 1.30E+08 0.026 0.013
4 6 10 9 0 1.20E+07 2.88E+08 0.021 0.01
4 7 10 10 0 1.36E+07 3.81E+08 0.013 0.006
4 8 10 10 0 1.38E+07 4.42E+08 0.011 0.005

Notes: The number of replicates (out of 10) with a given number of chains (Ch) are shown which found both peaks (Peak), converged to an RMSD at most 0.02 (Conv), or exceeded the iteration limit (Lim). The mean number of iterations, running time (iterations*chains*runs), maximum split frequency error (MaxErr), and RMSD are also shown.