vivarium-comparelisted

# vivarium-compare — genome relatedness, gene content, and structure Answer the three questions of comparative genomics — *how related?*, *what gene content do they share or not?*, *is the structure conserved?* — and return tables a reviewer can read, plus figures via vivarium-report. Light analyses run now in the `bio_tools` env; the one slow step (OrthoFinder) is handed back as an exact command rather than blocking the session. ## Step 1 — pick the comparison | Question | Analysis | Tool | Weight | |---|---|---|---| | How similar are these genomes? Same species? | **ANI** (nucleotide identity) | FastANI | light → run (bundled) | | Relatedness at the protein level (more sensitive, distant taxa) | **AAI** | EzAAI | light → run | | What genes are shared vs unique? Core/accessory? Pangenome? | **orthology** | OrthoFinder | heavy → **scaffold** | | Is gene order / structure conserved? Rearrangements? | **synteny** | MUMmer (nucmer) | moderate → run (bundled) | ANI ≥ ~95% is the conventional same-species boundary (Jain et al. 2018) — state it, don't assume it. ## Step 2 — run the light steps (bundled `compare.sh`) Call the script under this skill's directory by full path (Claude's cwd is the user's project). **ANI, all-vs-all → a square matrix** (drop it straight into vivarium-report `heatmap`): ```bash bash <skill-dir>/scripts/compare.sh ani --indir <genomes_dir> --out ani_matrix.tsv # or an explicit list: --list genomes.txt (one genome path per line) ``` The matrix