Output reference

Every command writes beneath --output (default ./). Aviary keeps working products in data/ and exposes important final products through stable top-level paths, often as symbolic links. The exact tree depends on the command, input types and skipped stages.

Common run directories

output_dir/
├── assembly/       final assembly link
├── bins/           final MAGs and genome summaries
├── annotation/     functional annotation products
├── taxonomy/       raw taxonomic classification products
├── diversity/      SingleM products when enabled
├── www/            quality-control reports and summaries
├── benchmarks/     per-rule Snakemake benchmark records
├── logs/           per-rule standard output and error logs
└── data/           working data and canonical generated files

Do not assume every directory is present. For example, assemble does not produce bins/, and --skip-taxonomy suppresses the corresponding taxonomy stage.

Assembly outputs

Path	Format	Status	Meaning
assembly/final_contigs.fasta	FASTA	final link	Assembly intended for downstream use
data/final_contigs.fasta	FASTA	final, canonical	Generated assembly targeted by the stable link
www/assembly_stats.txt	text	final report	Assembly size statistics
www/fastp.html	HTML	report, when short-read QC runs	Interactive short-read filtering report
www/rastqc/	HTML/support files	report, when enabled	Short-read contamination/QC report
www/rastqc_long/	HTML/support files	report, when enabled	Long-read contamination/QC report
www/nanoplot/	HTML/support files	report, with long reads	Long-read quality summary

Genome recovery outputs

Path	Format	Status	Meaning
bins/final_bins/	directory of FASTA	final	Recovered MAG sequences
bins/bin_info.tsv	TSV	final	Main per-MAG quality, taxonomy and assembly summary
bins/checkm_minimal.tsv	TSV	final	Compact CheckM2 quality summary
bins/coverm_abundances.tsv	TSV	final when abundance runs	CoverM abundance estimates by sample
taxonomy/	directory	final/supporting	Raw GTDB-Tk results
diversity/metagenome.combined_otu_table.csv	CSV	final when SingleM runs	Read-based community profile
diversity/singlem_appraisal.tsv	TSV	final when SingleM appraisal runs	Comparison of read and genome recovery evidence
diversity/singlem_appraise.svg	SVG	final figure	Visual SingleM appraisal summary

Treat bin_info.tsv as a summary, not a substitute for retaining the raw quality and taxonomy outputs. Column sets can depend on enabled stages; inspect the header from the installed version before writing downstream parsers.

Annotation outputs

Functional and taxonomic products are written below annotation/ and taxonomy/ according to the selected workflow targets. Preserve their logs and database release information when interpreting or publishing results.

Logs and benchmarks

Files below logs/ are the first place to investigate a failed rule. Files below benchmarks/ are Snakemake benchmark records for completed rules and can support local resource planning. They are observations from the current input and environment, not general performance guarantees.

Intermediate files and cleanup

The data/ directory contains both canonical generated products and working files. Do not remove it wholesale. With --clean enabled (the default), Snakemake removes files declared temporary after their consumers finish. Use --clean false only when retaining those intermediates is worth the storage cost or they are needed for a planned partial workflow.