sequana-lora 1.2.0

PacBio HiFi (recommended setup)

sequana_lora \
    --input-directory /data/hifi \
    --data-type pacbio-hifi \
    --assembler flye \
    --genome-size 3m \
    --apptainer-prefix /shared/containers \
    --do-coverage

Nanopore (bacteria, full quality pipeline)

Use --mode bacteria to enable sequana_coverage, prokka, busco, and checkm in one shot:

sequana_lora \
    --input-directory /data/nanopore \
    --data-type nano-hq \
    --assembler flye \
    --genome-size 3m \
    --apptainer-prefix /shared/containers \
    --mode bacteria \
    --busco-lineage bacteria \
    --checkm-rank genus \
    --checkm-name Streptococcus

PacBio subreads (with CCS construction)

If your input is raw PacBio BAM files (subreads), LORA can build CCS/HiFi reads first:

sequana_lora \
    --input-directory /data/subreads \
    --data-type pacbio-raw \
    --assembler flye \
    --genome-size 3m \
    --pacbio-build-ccs \
    --pacbio-ccs-min-passes 10 \
    --pacbio-ccs-min-rq 0.99

Or, if you have multiple BAM files per sample, provide a CSV:

sequana_lora \
    --pacbio-input-csv samples.csv \
    --data-type pacbio-raw \
    --assembler flye \
    --genome-size 3m

The CSV format is: one row per sample with columns sample,file1[,file2,...].

Coverage analysis

Computes depth of coverage and breadth of coverage for each contig using sequana_coverage. Highly recommended to check assembly quality:

--do-coverage

BUSCO completeness

Assess genome completeness against a lineage-specific marker gene set:

--busco-lineage bacteria          # auto-download bacteria lineage
--busco-lineage streptococcales   # specific clade
--busco-print-lineages            # list all available lineages

Prokka annotation

Annotate contigs (bacterial genomes):

--do-prokka

CheckM genome quality

Estimate completeness and contamination for bacterial genomes:

--checkm-rank genus --checkm-name Streptococcus

Use an invalid --checkm-name value to get a list of valid names for a given rank, e.g. --checkm-rank genus --checkm-name HELP.

Polypolish (Illumina polishing)

Polish long-read contigs with paired-end Illumina data:

--do-polypolish \
--polypolish-input-directory /data/illumina \
--polypolish-input-pattern "*.fastq.gz" \
--polypolish-input-readtag "_R[12]_"

Circularisation

Explicit circularisation with Circlator (Flye performs this automatically):

--do-circlator

Local BLAST

Requires a locally installed BLAST+ and a downloaded nt database (~270 GB). Fastest option with no network dependency:

--blastdb /path/to/blast/databases

Remote BLAST (NCBI)

No local database required — jobs are submitted to NCBI’s BLAST servers. Enable by providing an email address:

--blast-email your@email.com

Jobs are submitted sequentially (one contig at a time) to avoid IP-level CPU throttling by NCBI. The default database is nt; use --blast-remote-db to change it:

--blast-email your@email.com --blast-remote-db refseq_genomic

Restricting the search to an organism group (strongly recommended)

Searching all of nt for a large contig is slow and prone to NCBI CPU throttling. Restrict the search by editing config.yaml after running sequana_lora. The entrez_query parameter is equivalent to filling the “Organism” box on the NCBI BLAST web form.

Option 1 — curated bacterial reference genomes (fastest, recommended for bacteria)

Use refseq_genomic as the database and restrict to bacteria. RefSeq contains only complete, curated reference genomes — a much smaller and higher-quality search space than nt:

blast:
    remote_db: 'refseq_genomic'
    entrez_query: 'Bacteria[Organism]'

Option 2 — all bacteria in nt, reference sequences only

Stay on nt but filter to RefSeq-quality entries:

blast:
    remote_db: 'nt'
    entrez_query: 'Bacteria[Organism] AND refseq[filter]'

Option 3 — single genus

Useful when the organism is known:

blast:
    remote_db: 'nt'
    entrez_query: 'Streptococcus[Organism]'

NCBI API key (optional but recommended)

Register for a free NCBI API key at https://www.ncbi.nlm.nih.gov/account/ (sign in → API Key Management). It raises the rate limit from 3 to 10 requests/second and reduces CPU throttling for large queries. Add it to config.yaml:

blast:
    api_key: 'YOUR_KEY_HERE'