Metagenome assembly

Assembly transforms quality-controlled reads into contigs for genome recovery or other downstream analysis. Aviary supports short-read, long-read and hybrid inputs and chooses applicable workflow rules from the data supplied.

Select an input strategy

• Paired short reads require matching -1 and -2 lists.
• Long reads require --longreads and a --long-read-type value.
• Hybrid assembly supplies both short and long reads.
• Multiple samples can be co-assembled where the selected options support it.

Input quality control happens before assembly unless it is explicitly skipped. Host references supplied through --host-filter are used to remove matching reads before the assembly stage.

Basic usage

aviary assemble \
  -1 sample_R1.fastq.gz \
  -2 sample_R2.fastq.gz \
  --output assembly_run \
  --max-threads 16 \
  --n-cores 16

For a hybrid dataset, add long reads and their platform type:

aviary assemble \
  -1 sample_R1.fastq.gz \
  -2 sample_R2.fastq.gz \
  --longreads sample_ont.fastq.gz \
  --long-read-type ont \
  --output hybrid_run

Important controls

--long-read-assembler selects the supported long-read assembly implementation. Quality-control thresholds affect which reads reach the assembler and should be chosen for the sequencing data, not copied mechanically between projects. --min-contig-size is accepted by assemble through its shared parser, but it is a downstream binning filter and does not alter the default assembly target; set it on recover or complete when filtering contigs for binning.

Result

The canonical assembly is stored in data/final_contigs.fasta and exposed as assembly/final_contigs.fasta. Reports are written under www/; rule logs and benchmarks are kept in logs/ and benchmarks/.

See the assemble reference for every option and the output reference for the run layout.