IQ2MC: A New Framework to Infer Phylogenetic Time Trees Using IQ-TREE 3 and MCMCtree with Mixture Models

This contains raw data for the manuscript “IQ2MC: A New Framework to Infer Phylogenetic Time Trees Using IQ-TREE 3 and MCMCtree with Mixture Models” Empirical_data.zip: multiple sequence alignments for each analyzed empirical alignment, calibrated tree files, partition files, and input control files for MCMCTree. Empirical_data_analysis.zip: scripts and MCMCTree output files for comparing divergence times under various evolutionary models and four calibration settings: young, middle, root, and all. Simulation_data.zip: Simulated DNA and amino acid alignments, the tree topologies they were generated on, and the analysis outputs for the simulation sections of the study. Alignments were generated with AliSim-HPC under a strict clock on four topologies — two unbalanced (caterpillar) trees and two birth-death trees, each with 16 and 64 taxa, root age fixed at 3 time units — under six evolutionary rate categories: 1x, 2x, 4x, 6x, 8x, and 10x, corresponding to 0.1, 0.2, 0.4, 0.6, 0.8 and 1.0 substitutions per site per unit time. DNA data use the GTR+G4 substitution model, while AA data use LG+G4 and six profile mixture models (LG+G4+C10 through LG+G4+C60). All alignments have 5,000 sites and a fixed gamma shape of α = 0.25. Each was analyzed under three fossil calibration schemes — young, middle and root — which differ in where in the tree the most informative calibration is placed; under the young and middle schemes the root carries a diffuse gamma calibration, as MCMCtree requires a calibration on the root. trees: The four tree topologies, each under the three calibration schemes (12 files). Branch lengths are in time units and become substitution units when multiplied by the rate category. model_misspecification: The main simulation grid — 24 DNA alignments (4 topologies × 6 rate categories) and 168 amino acid alignments (4 topologies × 6 rate categories × 7 substitution models), organized by topology. Alignments only; the corresponding trees are in trees/. The DNA data were analyzed under the true GTR+G4 model and the deliberately misspecified JC69 model; the amino acid data simulated under profile mixture models were analyzed under the true mixture model and under the simpler LG+G4 model. PMSF_simulation: Simulations in which every site has its own empirical amino acid profile, estimated with the posterior mean site frequency (PMSF) method and taken from Giacomelli et al. (2025). Alignments of 500, 1,000, 2,000, and 5,000 sites were simulated under the LG+G4+PMSF model on the four topologies at two rate categories (1x and 10x), in five replicates, and dated under LG+G4 and LG+G4+C60. Contains the four site-profile sets, the 160 simulated alignments, and the calibrated trees used for the simulations. approximate_likelihood: Data behind the assessment of how faithfully MCMCtree's second-order Taylor approximation reproduces the exact phylogenetic likelihood. Two experiments. /two_taxa profiles the exact log likelihood along a single branch of a three-taxon tree, contrasting a branch whose maximum likelihood estimate lies on the zero boundary against interior optima at short and long branch lengths; it holds the two simulated 500-site alignments, their trees, the gradient and Hessian files, and the exact log likelihoods along each profiled branch. /asymtotic measures how the error of the approximation decays as the alignment grows, on 16-taxon unbalanced and birth-death trees, using amino acid alignments of 500 to 1,000,000 sites; it holds the 16 alignments, the calibrated trees, and the summary figures of relative error and rank correlation against alignment length. Every folder carries a README.txt with the naming conventions and details of that experiment. Note that the analysis pipelines write the three calibration schemes as "shallow", "mid" and "old"; these correspond to young, middle and root respectively. The empirical data were derived from the following sources: Álvarez-Carretero, Sandra, Asif U. Tamuri, Matteo Battini, Fabrícia F. Nascimento, Emily Carlisle, Robert J. Asher, Ziheng Yang, Philip C. J. Donoghue, and Mario Dos Reis. 2022. “A Species-Level Timeline of Mammal Evolution Integrating Phylogenomic Data.” Nature 602 (7896): 263–67. Betts, Holly C., Mark N. Puttick, James W. Clark, Tom A. Williams, Philip C. J. Donoghue, and Davide Pisani. 2018. “Integrated Genomic and Fossil Evidence Illuminates Life’s Early Evolution and Eukaryote Origin.” Nature Ecology & Evolution 2 (10): 1556–62. Morris, Jennifer L., Mark N. Puttick, James W. Clark, Dianne Edwards, Paul Kenrick, Silvia Pressel, Charles H. Wellman, Ziheng Yang, Harald Schneider, and Philip C. J. Donoghue. 2018. “The Timescale of Early Land Plant Evolution.” Proceedings of the National Academy of Sciences of the United States of America 115 (10): E2274–83. Reis, Mario dos, Yuttapong Thawornwattana, Konstantinos Angelis, Maximilian J. Telford, Philip C. J. Donoghue, and Ziheng Yang. 2015. “Uncertainty in the Timing of Origin of Animals and the Limits of Precision in Molecular Timescales.” Current Biology: CB 25 (22): 2939–50. The PMSF profiles were extracted for the PMSF simulations from: Giacomelli, Mattia, Matteo Vecchi, Roberto Guidetti, et al. 2025. “CAT-Posterior Mean Site Frequencies Improves Phylogenetic Modeling under Maximum Likelihood and Resolves Tardigrada as the Sister of Arthropoda plus Onychophora.” Genome Biology and Evolution 17 (1): evae273.