The following data archives are provided: ########## ef2.tar.gz: ########## ef2.ali: a multiple sequence alignment of elongation factor 2 in 30 eukaryotic species (627 aligned positions), taken from Lartillot and Philippe, 2006. Computing Bayes factor using thermodynamic integration. Syst Biol 55:195. phylip format ########## metazoa.tar.gz ########## meta.ali: a concatenation of genes (35371 aligned positions) across 35 metazoans + 2 choanoflagellates and 12 fungi, originally published in Philippe et al, 2005. Multigene analyses of bilaterian animals corroborate the monophyly of Ecdysozoa, Lophotrochozoa and Protostomia, Mol Biol Evol 22:1246 phylip format meta80.ali: the meta.ali alignment was filtered to remove sites with more than 20% missing data, leaving a total of 9804 sites. ########## table1_simu_replicates.tar.gz ########## the folder in this directory contains the data replicates simulated under the multivariate normal model, which were used for computing Table 1. These simulated data were produced using the normcv package (see normcv/data/README.md for more details) ########## table2_cv_replicates.tar.gz ########## cross-validation replicates for ef2.ali This directory contains all data replicates that were used for cross-validation on the elongation factor 2 data set (table 2). These have been produced using the original alignment: ef2.ali for each replicate rep=0..29: - cvef2${rep}.ali : randomly re-shuffled version of the original data (5-fold CV by SMC) - cvef2${rep}_learn.ali : the first (1-f)*nsite sites of cvef2${rep}.ali: for training (5-fold CV by naive IS) - cvef2${rep}_test.ali : the last f*nsite sites of cverf${rep}.ali: for validation (5-fold CV by naive IS) Note that a script for producing similar replicates is provided in the directory for reproducing the results of Table 2 (table2_ef2) ########## figure2_jackknife_replicates.tar.gz ########## This folder contains all jackknife replicates (based on the original empirical alignment meta80.ali) that were used for computing the LOO-CV and Bayes factor scores of Figure 2 naming convention: subfullmeta{nsite}_{rep}.ali, for nsite=100..800 and rep=0..9 A script for drawing similar jackknife replicates is given in the directory for reproducing Figures 2 and 3 (figures2_3_meta_lg_gtr) ########## figure3_jackknife_replicates.tar.gz ########## This directory contains: - lgmeta_ppred{rep}.ali, for rep=1,2,3,4: the simulation replicates obtained by posterior predictive simulations (by running the LG model on meta80.ali) - ppredlgsubmeta{nsite}_{rep}.ali, for nsite=100...800 and rep=0,1,2,3: the jacknife replicates that were produced based on these simulated data for computing the LOO-CV and Bayes factor scores of Figure 3 A script for drawing similar jackknife replicates is given in the directory for reproducing Figures 2 and 3 (figures2_3_meta_lg_gtr) ########## figure4_jackknife_replicates.tar.gz ########## This directory contains all jackknife replicates (based on the original empirical alignment meta.ali) that were used for conducting the scaling experiment shown in Figure 4. names follow this convention: subfullmeta{nsite}_{rep}.ali for nsite=500, 1000, 2000, 4000, 8000, 16000 and rep=0, 1, 2, 3