COMEBin metagenomic binning benchmark — datasets and run results

Benchmark of metagenomic genome binners (starting with COMEBin v1.1.0) with a repeatable harness: per-run parameters, datasets, wall time, and bin quality scored with CheckM2 and CheckM (v1). Repository: https://github.com/paulzierep/metagenomic-binning-benchmark (setup, datasets, evaluation and run docs under docs/; resume state in PROGRESS.md). This record accumulates ALL newly generated benchmark data that is not publicly available otherwise, as continuous versions under the concept DOI 10.5281/zenodo.22935024 (each new evaluated run publishes a new version). Version 9.0 (this version) contains the benchmark results of every evaluated run: baseline_rerun_autorestart1 (60 bins) cami3_v11_20260926_rerun (492 bins) fix_v11_20260924 (71 bins) human_v11_20260925 (107 bins) marine_v11_20260925 (152 bins) medium_v11_20260924 (16 bins) small_test_v4 (3 bins) sweep_001_ref (17 bins) sweep_003_issue28fix (17 bins) sweep_004_temp005 (13 bins) sweep_005_temp030 (17 bins) sweep_006_temp050 (17 bins) sweep_007_emb1024 (15 bins) tiny_test_n101 (1 bins). Each run contributes its bin FASTAs (runs//comebin_res/comebin_res_bins/), per-bin and aggregate CheckM2 + CheckM v1 metrics (results/.csv, runs//per_bin_results.csv), the evaluation reports (runs//eval/), run parameters and logs (run_meta.txt, comebin_run.log), the CAMI-standard binning export (runs//biobox/: binning.tar.gz, .binning in CAMI binning format v0.9.1, binning_summary.tsv), all comparison figures (results/figures/, incl. the combined completeness bar plot), plus dataset, environment, fix, evaluation and preservation documentation (docs/). manifest.json inside the archive lists every file with size and SHA-256. Version 1.0 additionally contains the small functional-test dataset (comebin_small_release_v1.tar.gz, 300 contigs + overlapping reads). How to run: clone the repository, follow docs/02/06 to build the environment, run a dataset with scripts/run_comebin_fix.sh and evaluate with scripts/run_eval.sh; docs/01-datasets.md and docs/11-run-results.md describe each dataset and run.