First complete analysis¶
aviary complete joins assembly, genome recovery and annotation into one
Snakemake run. Use it when you want Aviary to carry raw reads through to a
curated set of metagenome-assembled genomes (MAGs).
Data flow¶
paired short reads ─┐
├─ quality control ─ assembly ─ read mapping
optional long reads ┘ │
▼
binning ─ refinement ─ quality assessment
│
▼
taxonomy and functional annotation
Aviary orchestrates these stages; specialised dependencies perform the underlying assembly, binning, quality assessment and annotation algorithms.
Choose the input mode¶
| Data available | Required input | Typical use |
|---|---|---|
| Paired short reads | -1, -2 |
Short-read metagenome assembly and recovery |
| Long reads | --longreads, --long-read-type |
Long-read assembly and recovery |
| Both | short- and long-read options | Hybrid assembly and recovery |
| Existing assembly | --assembly plus reads for coverage |
Skip de novo assembly and recover MAGs |
See the input reference for pairing rules and accepted long-read type identifiers.
Run in a dedicated output directory¶
aviary complete \
-1 reads/sample_R1.fastq.gz \
-2 reads/sample_R2.fastq.gz \
--longreads reads/sample_ont.fastq.gz \
--long-read-type ont \
--output sample_aviary \
--max-threads 16 \
--n-cores 32 \
--max-memory 250
The default maximum memory value in the workflow configuration is 250 GB. Set the option to the actual hard limit available to the run; it is not a prediction of typical memory consumption.
Monitor and resume¶
Rule-specific messages are written beneath logs/, while Snakemake benchmark
records are written beneath benchmarks/. Re-running the same command and
output directory normally resumes from existing valid outputs. Do not delete
.snakemake/ or intermediate files while diagnosing an interrupted run.
For scheduler submission, resource caps and retries, see
HPC and cluster submission. For precise command options, see
the complete CLI reference.