Cohort-HMM marker recruitment with per-OG orthology QC for phylogenomic supermatrices — code and per-case data products

Reproducibility deposit for the Cohort-HMM marker recruitment paper (Nielsen 2026, bioRxiv [DOI TBD]). Contains the paper draft (Markdown + PDF), all pipeline scripts (Pyrodigal → OrthoFinder → HMMER → MAFFT → BMGE2 → IQ-TREE), and per-case data products for three case studies spanning genus, family, and phylum ranks: Actinomarina (acti/), Pelagibacteraceae (pela/), and Omnitrophota (omni/). For each case the deposit includes input proteomes (inhouse cohort + NCBI species representatives from GTDB R232), per-OG profile HMMs and hmmsearch tables, per-OG alignments and gene trees, the concatenated supermatrix, the IQ-TREE result, and the orthology-QC and recovery TSVs. The pre-computed OrthoFinder all-vs-all output is omitted; Orthogroups.txt and the per-OG HMMs are sufficient for downstream re-use. See AGENT_GUIDE.md for the directory map, per-case pipeline, and conventions.