IQ-TREE3-GPU benchmark data: source genomes, alignments and the paired CPU-GPU run record

Input data and measured record of the benchmark in the manuscript IQ-TREE3-GPU: GPU-accelerated maximum-likelihood phylogenetic tree inference for complex amino acid substitution models (Byeongyeon Cho and Aleksandar D. Kostic; corresponding author Aleksandar D. Kostic, [e-mail]). Companion to the IQ-TREE3-GPU software repository, https://github.com/kosticlab/iqtree3-gpu (mirror https://github.com/terry-b-cho/iqtree3-gpu), which holds the tool, the benchmark harness, the analysis and figure code and the documentation. This deposit lets the published results be reproduced without Git LFS and makes the processed datasets available for reuse. Contents. benchmark_data_genomes: 435 publicly deposited bacterial metagenome-assembled genomes (14 GTDB genera, eight studies) with a per-genome metadata table (study, public accessions, host, GTDB-Tk taxonomy, CheckM2 quality). benchmark_data_protein: 45 concatenated marker-gene amino-acid alignments built from those genomes (Prodigal, DIAMOND against the PhyloPhlAn 3 marker database, MAFFT), 35 of which are the benchmark datasets of the manuscript. benchmark_data_dna: 26 nucleotide alignments simulated with AliSim in IQ-TREE 3, with guide trees, seeds and logs. benchmark_runs: the paired CPU-versus-GPU run record (RUNS.tsv with one row per paired run, phase timings, replicate table, key numbers, every run directory with both arms' IQ-TREE logs, trees, support files and equivalence results, and the column dictionary). supporting_files: README.md explaining every folder with full citations, MANIFEST.sha256 and ARCHIVES.sha256 checksums, BENCHMARK_SCHEMA.md and install_into_repo.sh, which unpacks the archives into a clone of the repository. Credit for the genomes. The genomes are copies of assemblies deposited by Li et al. 2025 (160 genomes), Sanders et al. 2023 (151), Carter et al. 2023 (78), Bowerman et al. 2021 (21), Glendinning et al. 2021 (14), Svartström et al. 2017 (8), Pu et al. 2023 (2) and Wibowo et al. 2021 (1), renamed to stable identifiers; the metadata table maps each file to its study and public accessions. Users of the genomes should cite the study that produced them (listed under related identifiers and in README.md); the genomes remain subject to the terms of the archives they were obtained from. Licence. The compilation is released under CC BY 4.0. IQ-TREE 3 (Wong et al. 2026, doi:10.1093/molbev/msag117) must be cited alongside IQ-TREE3-GPU. Tree of this deposit. Archive names are the repository paths with / replaced by _; each unpacks to that path inside a clone of the repository. doi:10.5281/zenodo.22942159/ |-- benchmark_data_genomes.tar.gz (262 MB) unpacks to benchmark/data/genomes/ | |-- PROVENANCE.md | |-- genome_metadata.tsv | |-- study_citations.tsv | |-- g__Brotaphodocola/ (10 genomes, <Host_species>_<lifestyle>_<study>_<id>.fna) | |-- g__CAG-269/ (10 genomes, <Host_species>_<lifestyle>_<study>_<id>.fna) | |-- g__Caccovivens/ (17 genomes, <Host_species>_<lifestyle>_<study>_<id>.fna) | |-- g__Colicola/ (65 genomes, <Host_species>_<lifestyle>_<study>_<id>.fna) | |-- g__Coprococcus_C/ (30 genomes, <Host_species>_<lifestyle>_<study>_<id>.fna) | |-- g__Coprovivens/ (13 genomes, <Host_species>_<lifestyle>_<study>_<id>.fna) | |-- g__Enterousia/ (54 genomes, <Host_species>_<lifestyle>_<study>_<id>.fna) | |-- g__Hominicoprocola/ (12 genomes, <Host_species>_<lifestyle>_<study>_<id>.fna) | |-- g__Lachnospira/ (25 genomes, <Host_species>_<lifestyle>_<study>_<id>.fna) | |-- g__Nanosyncoccus/ (20 genomes, <Host_species>_<lifestyle>_<study>_<id>.fna) | |-- g__Onthovivens/ (34 genomes, <Host_species>_<lifestyle>_<study>_<id>.fna) | |-- g__Roseburia/ (41 genomes, <Host_species>_<lifestyle>_<study>_<id>.fna) | |-- g__Ruminococcoides/ (66 genomes, <Host_species>_<lifestyle>_<study>_<id>.fna) | `-- g__Ruminococcus/ (38 genomes, <Host_species>_<lifestyle>_<study>_<id>.fna) |-- benchmark_data_protein.tar.gz (55 MB) unpacks to benchmark/data/protein/ | |-- README.md | |-- 1Thread_g__Brotaphodocola_n6/ (PROVENANCE.md, concat_alignment.faa) | |-- 1Thread_g__CAG-269_n6/ (PROVENANCE.md, concat_alignment.faa) | |-- 1Thread_g__Caccovivens_n6/ (PROVENANCE.md, concat_alignment.faa) | |-- 1Thread_g__Colicola_n6/ (PROVENANCE.md, concat_alignment.faa) | |-- 1Thread_g__Coprococcus_C_n6/ (PROVENANCE.md, concat_alignment.faa) | |-- 1Thread_g__Coprovivens_n6/ (PROVENANCE.md, concat_alignment.faa) | |-- 1Thread_g__Enterousia_n6/ (PROVENANCE.md, concat_alignment.faa) | |-- 1Thread_g__Hominicoprocola_n6/ (PROVENANCE.md, concat_alignment.faa) | |-- 1Thread_g__Lachnospira_n6/ (PROVENANCE.md, concat_alignment.faa) | |-- 1Thread_g__Nanosyncoccus_n6/ (PROVENANCE.md, concat_alignment.faa) | |-- 1Thread_g__Onthovivens_n6/ (PROVENANCE.md, concat_alignment.faa) | |-- 1Thread_g__Roseburia_n6/ (PROVENANCE.md, concat_alignment.faa) | |-- 1Thread_g__Ruminococcoides_n6/ (PROVENANCE.md, concat_alignment.faa) | |-- 1Thread_g__Ruminococcus_n6/ (PROVENANCE.md, concat_alignment.faa) | |-- g__Brotaphodocola/ (concat_alignment.faa, tree_qc.tsv) | |-- g__CAG-269/ (concat_alignment.faa, tree_qc.tsv) | |-- g__Caccovivens/ (concat_alignment.faa, tree_qc.tsv) | |-- g__Colicola/ (concat_alignment.faa, tree_qc.tsv) | |-- g__Coprococcus_C/ (concat_alignment.faa, tree_qc.tsv) | |-- g__Coprovivens/ (concat_alignment.faa, tree_qc.tsv) | |-- g__Enterousia/ (concat_alignment.faa, tree_qc.tsv) | |-- g__Hominicoprocola/ (concat_alignment.faa, tree_qc.tsv) | |-- g__Lachnospira/ (concat_alignment.faa, tree_qc.tsv) | |-- g__Nanosyncoccus/ (concat_alignment.faa, tree_qc.tsv) | |-- g__Onthovivens/ (concat_alignment.faa, tree_qc.tsv) | |-- g__Roseburia/ (concat_alignment.faa, tree_qc.tsv) | |-- g__Ruminococcoides/ (concat_alignment.faa, tree_qc.tsv) | |-- g__Ruminococcus/ (concat_alignment.faa, tree_qc.tsv) | |-- n04_g__Brotaphodocola/ (concat_alignment.faa) | |-- n06_g__CAG-269/ (concat_alignment.faa) | |-- n08_g__Hominicoprocola/ (concat_alignment.faa) | |-- n10_g__Coprovivens/ (concat_alignment.faa) | |-- n12_g__Caccovivens/ (concat_alignment.faa) | |-- n14_g__Nanosyncoccus/ (concat_alignment.faa) | |-- n17_g__Lachnospira/ (concat_alignment.faa) | |-- n20_g__Coprococcus_C/ (concat_alignment.faa) | |-- n23_g__Onthovivens/ (concat_alignment.faa) | |-- n26_g__Ruminococcus/ (concat_alignment.faa) | |-- n28_g__Roseburia/ (concat_alignment.faa) | |-- n30_g__Enterousia/ (concat_alignment.faa) | |-- p_lc_n04_L200/ (PROVENANCE.md, concat_alignment.faa) | |-- p_lc_n04_L20k/ (PROVENANCE.md, concat_alignment.faa) | |-- p_lc_n04_L2k/ (PROVENANCE.md, concat_alignment.faa) | |-- p_lc_n06_L20k/ (PROVENANCE.md, concat_alignment.faa) | `-- p_lc_n08_L50k/ (PROVENANCE.md, concat_alignment.faa) |-- benchmark_data_dna.tar.gz (755 MB) unpacks to benchmark/data/dna/ | |-- CHECKSUMS.sha256 | |-- README.md | |-- TRUE_MODEL.txt | |-- dna_hc_n100_L2M/ (PROVENANCE.md, alisim.log, aln.phy, aln.treefile, aln.treefile.log) | |-- dna_hc_n30_L1M/ (PROVENANCE.md, alisim.log, aln.phy, aln.treefile, aln.treefile.log) | |-- dna_hc_n30_L3M/ (PROVENANCE.md, alisim.log, aln.phy, aln.treefile, aln.treefile.log) | |-- dna_hc_n60_L2M/ (PROVENANCE.md, alisim.log, aln.phy, aln.treefile, aln.treefile.log) | |-- dna_hc_n60_L4M/ (PROVENANCE.md, alisim.log, aln.phy, aln.treefile, aln.treefile.log) | |-- dna_n04/ (alisim.log, aln.phy, aln.treefile, aln.treefile.log) | |-- dna_n06/ (alisim.log, aln.phy, aln.treefile, aln.treefile.log) | |-- dna_n08/ (alisim.log, aln.phy, aln.treefile, aln.treefile.log) | |-- dna_n10/ (alisim.log, aln.phy, aln.treefile, aln.treefile.log) | |-- dna_n12/ (alisim.log, aln.phy, aln.treefile, aln.treefile.log) | |-- dna_n14/ (alisim.log, aln.phy, aln.treefile, aln.treefile.log) | |-- dna_n17/ (alisim.log, aln.phy, aln.treefile, aln.treefile.log) | |-- dna_n20/ (alisim.log, aln.phy, aln.treefile, aln.treefile.log) | |-- dna_n23/ (alisim.log, aln.phy, aln.treefile, aln.treefile.log) | |-- dna_n26/ (alisim.log, aln.phy, aln.treefile, aln.treefile.log) | |-- dna_n28/ (alisim.log, aln.phy, aln.treefile, aln.treefile.log) | |-- dna_n30/ (alisim.log, aln.phy, aln.treefile, aln.treefile.log) | |-- dna_pt_n14/ (PROVENANCE.md, alisim.log, aln.phy) | |-- dna_pt_n17/ (PROVENANCE.md, alisim.log, aln.phy) | |-- dna_pt_n20/ (PROVENANCE.md, alisim.log, aln.phy) | |-- dna_pt_n23/ (PROVENANCE.md, alisim.log, aln.phy) | |-- dna_pt_n26/ (PROVENANCE.md, alisim.log, aln.phy) | |-- dna_pt_n28/ (PROVENANCE.md, alisim.log, aln.phy) | |-- dna_pt_n30/ (PROVENANCE.md, alisim.log, aln.phy) | |-- dna_ux_n100_L4M/ (PROVENANCE.md, alisim.log, aln.phy, aln.treefile, aln.treefile.log) | `-- dna_ux_n60_L6M/ (PROVENANCE.md, alisim.log, aln.phy, aln.treefile, aln.treefile.log) |-- benchmark_runs.tar.gz (5 MB) unpacks to benchmark/runs/ | |-- KEY_NUMBERS.md | |-- KEY_NUMBERS.tsv | |-- PHASE_STEPS.tsv | |-- REPLICATES_20T.tsv | |-- RUNS.csv | |-- RUNS.tsv | |-- RUNS_summary.csv | |-- RUNS_summary.tsv | |-- dna_cpu_reference/dna_cpu_reference.tsv | `-- paired/ (226 run directories named <dataset>__T<threads>__<card namespace>/, each holding:) | |-- PHASES.tsv | |-- RESULT.tsv | |-- cpu.contree | |-- cpu.iqtree | |-- cpu.log | |-- cpu.rc | |-- cpu.treefile | |-- equivalence.kv | |-- gpu.contree | |-- gpu.iqtree | |-- gpu.log | |-- gpu.rc | |-- gpu.treefile | `-- gpu.utilfull `-- supporting_files.tar.gz unpacks to supporting_files/ |-- README.md (this description, with installation steps and full citations) |-- install_into_repo.sh (unpacks the four archives into a clone of the repository) |-- MANIFEST.sha256 (SHA-256 of every file in the unpacked tree) |-- ARCHIVES.sha256 (SHA-256 of the four archives) |-- BENCHMARK_SCHEMA.md (column dictionary of RUNS.tsv) |-- zenodo_metadata.json (the metadata of this record) `-- zenodo_deposit_benchmark_data.py (the script that staged and uploaded this record)