Prior specification is an essential component of parameter estimation and model comparison in Approximate Bayesian computation (ABC). Oaks et al. present a simulation-based power analysis of msBayes and conclude that msBayes has low power to detect genuinely random divergence times across taxa, and suggest the cause is Lindley's paradox. Although the predictions are similar, we show that their findings are more fundamentally explained by insufficient prior sampling that arises with poorly chosen wide priors that critically undersample nonsimultaneous divergence histories of high likelihood. In a reanalysis of their data on Philippine Island vertebrates, we show how this problem can be circumvented by expanding upon a previously developed procedure that accommodates uncertainty in prior selection using Bayesian model averaging. When these procedures are used, msBayes supports recent divergences without support for synchronous divergence in the Oaks et al. data and we further present a simulation analysis that demonstrates that msBayes can have high power to detect asynchronous divergence under narrower priors for divergence time. Our findings highlight the need for exploration of plausible parameter space and prior sampling efficiency for ABC samplers in high dimensions. We discus potential improvements to msBayes and conclude that when used appropriately with model averaging, msBayes remains an effective and powerful tool.
table_S1 _01272013
Supplemental 1. Genbank accession numbers and species names of data used to test for synchronous divergence with msBayes versions 20120222 and 20120510 given 25 species-pairs. Each pair of species has the same genus name.
table_S2_01272013
Supplemental 2. Genbank accession numbers and species names of data used to test for synchronous divergence with msBayes versions 20120222 and 20120510 given 23 species-pairs. Each pair of species has the same genus name.
table_S3_01272013
Supplemental 3. Genbank accession numbers and species names of data used to test for synchronous divergence with msBayes versions 20120222 and 20120510 given 23 species-pairs. Each pair of species has the same genus name.
batch_table1_prior
batch file to be used with msBayes to generate simulated draws from priors M1 through M8. The values for prior bounds
upperTheta, lowerTheta, upperTau, and upperAncPopSize will have to be changed accordingly depending on which of the 8 prior models (M7 in this example). IMPORTANT, if running in multiple parallel jobs, make sure the start times are different by more than a second to assure different random number seeds. To generate each set of prior simulations (M7 in this example), use command line: cd $HOME/msbayes/trunk/src
./msbayes.pl -c batch_table1_prior -r 500000 -o prior_M7
scripts_table1_posterior_rejectionsample
see #comments
scripts_Fig1_posterior_rejectionsample
read #comments
HU_IB_batchNEWpars_psi18_ne0.5_taumax0.75
batch file for generating PODS (used within Script_Table2_HUIB_PODS_PARSnewPsi18_3M_HUIBPODStest_neuralnet_psi_omega_et_tol0.1_110812_ne0.5_taumax0.75.sh). By changing appropriate prior settings, this batch file can also be used for obtaining the prior sample (ie Reference Table) used to generate the PODS
Script_Table2_HUIB_PODS_PARSnewPsi18_3M_HUIBPODStest_neuralnet_psi_omega_et_tol0.1_110812_ne0.5_taumax0.75
script for generating PODS (Table 2)
HUIB_PODS_PARSnewPsi18_3M_HUIBPODStest_neuralnet_psi_omega_et_tol0.1_061512_ne0.5_taumax0.75
to be used for generating PODS (table 2). Used within shell script Script_Table2_HUIB_PODS_PARSnewPsi18_3M_HUIBPODStest_neuralnet_psi_omega_et_tol0.1_110812_ne0.5_taumax0.75.sh
HUIBObsTAIL_columnReduction_ne0.5_taumax0.75
used for generating PODS (Table 2). Used within shell script Script_Table2_HUIB_PODS_PARSnewPsi18_3M_HUIBPODStest_neuralnet_psi_omega_et_tol0.1_110812_ne0.5_taumax0.75.sh
HU_IB_psi18_BFCalculation_110812
R script for calculating Bayes Factors given the PODS data (Table 2)
batchPulse1_26
batch file for generating prior sample for estimates in Table S4 (data Table S1)
batchPulse2_23
batch file for generating prior sample for estimates in Table S4 (data Table S2)
batchPulse3_22
batch file for generating prior sample for estimates in Table S4 (data Table S3)
BirdCOI R scripts_psi3
R scripts for generating estimates in table S4