name: anvio description: > Anvi'o — interactive multi-omics platform for metagenomic assemblies, MAG refinement, visualization, and SNP/profile exploration. Use when: (1) Manual or human-guided bin refinement, (2) Interactive contig profiles across samples, (3) Linking MAGs with coverage/SNVs/metatranscriptomes, (4) Producing Anvi'o interactive HTML reports. Complements automated ensemble binning (basalt/metawrap/dastool); not a replacement for CheckM2/GUNC QC. license: MIT metadata: version: "1.0.1" metagenomics.category: "analysis-tools" metagenomics.tags: "anvio, visualization, MAG, refinement, interactive" metagenomics.upstream: "https://github.com/merenlab/anvio" metagenomics.stage: "binning" metagenomics.maturity: "reviewed" metagenomics.last-reviewed: "2026-10-24" metagenomics.command-status: "contigs-database/profile workflow statically reviewed; interactive refinement not executed here"

Anvi'o

Upstream: merenlab/anvio · https://anvio.org/

Citation

Eren, A. M. et al. Anvi'o: an advanced analysis and visualization platform for 'omics data. PeerJ 3, e1319 (2015). https://doi.org/10.7717/peerj.1319

See also docs/references.md.

Analytical thinking

Anvi'o is a platform, not a single binner: contig databases, profiling BAM coverage across samples, interactive interfaces for organizing contigs into bins, SNV profiling, and multi-omics linking. Use it when automated bins need human-guided refinement or when you must communicate contig-level evidence.

Typical place in this package:

assembly + multi-sample BAMs
  → anvi'o contig DB + profile DB
  → import automatic bins (MetaBAT2 / BASALT / …) optional
  → interactive refinement
  → export bins → checkm2 / gunc / drep / gtdbtk

Do not treat interactive aesthetics as QC — still run checkm2 / gunc / deepurify as appropriate.

Prerequisites

How to run

The BAM must have been mapped to the exact FASTA used for the contigs database. If contig names need simplification, reformat before mapping and retain the name map; do not simplify the FASTA and then profile an old BAM.

set -Eeuo pipefail
id="${1:?usage: build-anvio-profile.sh SAMPLE_ID CONTIGS BAM OUTPUT_DIR}"
contigs="${2:?usage: build-anvio-profile.sh SAMPLE_ID CONTIGS BAM OUTPUT_DIR}"
bam="${3:?usage: build-anvio-profile.sh SAMPLE_ID CONTIGS BAM OUTPUT_DIR}"
out="${4:?usage: build-anvio-profile.sh SAMPLE_ID CONTIGS BAM OUTPUT_DIR}"
threads="${THREADS:-16}"
[[ "$id" =~ ^[A-Za-z0-9][A-Za-z0-9._-]*$ ]] || exit 2
test -s "$contigs" && test -s "$bam" || exit 2
samtools quickcheck -v "$bam"
{ test -s "$bam.bai" || test -s "${bam%.bam}.bai"; } || exit 2
test ! -e "$out" || { echo "use a new output directory: $out" >&2; exit 2; }
mkdir -p "$out"
 
anvi-gen-contigs-database -f "$contigs" -o "$out/CONTIGS.db" -n "$id" \
  >"$out/gen-contigs.stdout.log" 2>"$out/gen-contigs.stderr.log"
anvi-run-hmms -c "$out/CONTIGS.db" -T "$threads" \
  >"$out/hmms.stdout.log" 2>"$out/hmms.stderr.log"
anvi-profile -i "$bam" -c "$out/CONTIGS.db" \
  -o "$out/PROFILE_${id}" -T "$threads" \
  >"$out/profile.stdout.log" 2>"$out/profile.stderr.log"

Profile each sample separately against the same contigs database, then merge the resulting profile databases with a release-compatible anvi-merge command.

Decision tree

Need interactive MAG refinement / multi-omics viz?
├─ Yes → anvio
├─ Fully automated catalogue → basalt ★ / metawrap / dastool
└─ Only QC metrics → checkm2 / gunc / mag-qc

Validation, recovery, and provenance

Require readable contigs/profile databases, completed logs, BAM/reference name agreement, and expected contig/read counts. Before and after manual refinement, export collections with stable bin IDs and rerun checkm2, gunc, taxonomy, and read-support checks. Preserve the original automatic bins and an explicit contig move/add/remove audit; screenshots alone are not reproducible evidence.

Do not overwrite databases after failed migrations or release changes. Record Anvi'o release, database schemas, dependency versions, FASTA/BAM checksums, commands, logs, profile/merge sample order, collection exports, refinement notes, and final-bin checksums. Interactive judgment is curator evidence, not experimental validation.

Related skills

basalt · metawrap · coverm · checkm2 · gunc · deepurify · visualization · tool-selection · mag-qc