Skip to content
BranchesTags

Repository files navigation

bgc_quast_ppl

A Nextflow pipeline for biosynthetic gene cluster (BGC) prediction and comparison.

It takes one or more genome assemblies, predicts BGCs with three tools (antiSMASH, DeepBGC, GECCO), and then compares the results with bgc-quast. Heavily inspired from nf-core/funcscan 3.0.0 framework.

bgc_quast_ppl pipeline diagram/DAG

1. What you need before you start

Requirement Notes
Nextflow 25.10.5 This exact version. For now the pipeline doesn't Run on the latest version of the Nextflow which is 26.04.
Docker Must be installed and running. All tools run inside containers, so you don't install antiSMASH/DeepBGC/GECCO yourself.
antiSMASH database (v8) Downloaded once, stored locally. See section 3.
DeepBGC database Downloaded once, stored locally. See section 3.
Genome assemblies Nucleotide FASTA files (.fasta or .fasta.gz).

GECCO needs no external database — its model ships inside the container.

2. Get the pipeline

git clone <YOUR_REPO_URL> bgc_quast_ppl
cd bgc_quast_ppl

Pin the Nextflow version in every terminal you run the pipeline from:

export NXF_VER=25.10.5

Tip: run that line once per terminal session, before any nextflow command.

3. Databases

You need two databases on disk before running. Point the pipeline at them with absolute paths (full paths starting from /), not relative ones.

  • antiSMASH v8 database → passed with --bgc_antismash_db
  • DeepBGC database → passed with --bgc_deepbgc_db

These are downloaded using each tool's own standard download command. The exact commands depend on the tool version, so follow the official instructions:

Once downloaded, note the full path to each, for example:

~/bgc_quast_ppl/db/antismash_db_v8
~/bgc_quast_ppl/db/deepbgc_db

You will pass these paths on the command line (shown below).

4. The input samplesheet

The input is a CSV file. The columns are:

Column Meaning
sample A unique name for the genome (no spaces).
fasta Path to the genome FASTA (.fasta or .fasta.gz).
type Only needed in compare-to-reference mode. q = query, r = reference. Case does not matter (q/Q, r/R).

For compare-samples and compare-tools

The type column is not needed. A two-column sheet is enough:

sample,fasta
assembly_10,~/bgc_quast_ppl/data/assembly_10.fasta.gz
assembly_20,~/bgc_quast_ppl/data/assembly_20.fasta.gz

For compare-to-reference

You must include the type column, with exactly one reference row (r) and one or more query rows (q):

sample,fasta,type
assembly_10,~/bgc_quast_ppl/data/assembly_10.fasta.gz,q
assembly_20,~/bgc_quast_ppl/data/assembly_20.fasta.gz,q
reference,~/bgc_quast_ppl/data/reference.fasta.gz,r

The reference row's sample name (here reference) is used as the reference label in the bgc-quast report.

5. The three modes

You pick the mode with --bgc_quast_mode. Choose one per run.

Mode What it compares When to use it
compare-samples (default) The same tool across all your samples. You get one report per tool. "How do my genomes compare to each other?"
compare-tools The three tools against each other, per sample. You get one report per sample. "For this genome, how do antiSMASH, DeepBGC, and GECCO differ?"
compare-to-reference Your query genomes against one reference genome. QUAST runs automatically. "How do my genomes' BGCs compare to a known reference?"

6. How to run

Run from inside the bgc_quast_ppl directory. Replace the paths with your own.

Compare-samples (default)

export NXF_VER=25.10.5

nextflow run . \
  -profile docker \
  --input data/samplesheet.csv \
  --outdir results \
  --bgc_quast_mode compare-samples \
  --bgc_antismash_db ~/bgc_quast_ppl/db/antismash_db_v8 \
  --bgc_deepbgc_db ~/bgc_quast_ppl/db/deepbgc_db \
  --max_cpus 4 \
  --max_memory 24.GB

Compare-tools

Same command, change the mode:

  --bgc_quast_mode compare-tools \

Compare-to-reference

Use a samplesheet that has the type column and one reference row (section 4):

export NXF_VER=25.10.5

nextflow run . \
  -profile docker \
  --input data/samplesheet_ref.csv \
  --outdir results \
  --bgc_quast_mode compare-to-reference \
  --bgc_antismash_db ~/bgc_quast_ppl/db/antismash_db_v8 \
  --bgc_deepbgc_db ~/bgc_quast_ppl/db/deepbgc_db \
  --max_cpus 4 \
  --max_memory 24.GB

QUAST runs on its own in this mode — you do not set it up.

7. Common options

All optional. Defaults are shown.

Resources

Option Default Meaning
--max_cpus 2 Most CPUs any single step may use.
--max_memory 8.GB Most memory any single step may use.
--max_time 24.h Time limit per step.

Set these to fit your machine (e.g. --max_cpus 4 --max_memory 24.GB).

Choosing tools

Option Effect
--bgc_skip_antismash Skip antiSMASH.
--bgc_skip_deepbgc Skip DeepBGC.
--bgc_skip_gecco Skip GECCO.

By default all three run.

antiSMASH depth

antiSMASH runs in minimal mode by default (faster, core BGC detection only).

Option Effect
--bgc_antismash_full Run the full analysis instead of minimal.
--bgc_antismash_minimal Force minimal (this is already the default).

Do not pass both at once — the pipeline stops with an error if you do.

bgc-quast tuning

Option Default Meaning
--bgc_quast_edge_distance 100 Distance (bp) from a contig edge used to call a BGC "incomplete".
--bgc_quast_min_bgc_length 0 Ignore BGCs shorter than this. 0 = no minimum.
--bgc_quast_merge_distance 0 Merge BGCs closer than this.
--bgc_quast_overlap_fraction 0.9 Overlap fraction used when matching BGCs.
--bgc_quast_output_bgcs false Also write the individual BGC sequences.
--bgc_quast_quastdir — Supply your own QUAST output directory (compare-to-reference only). QUAST is then skipped.
--bgc_quast_debug false Print the raw error output on failure (for troubleshooting).

BiG-SCAPE (gene cluster families)

Optional. Runs BiG-SCAPE on the antiSMASH predictions and adds gene cluster family (GCF) rows plus a Venn diagram to the antiSMASH bgc-quast report. compare-samples mode only. Off by default — with it off, the report is unchanged.

Option Default Meaning
--run_bigscape false Turn it on. Needs antiSMASH enabled.
--bgc_bigscape_pfam — Path to an existing Pfam-A.hmm file (not its folder). Leave unset to download it automatically.
--bgc_bigscape_pfam_url Pfam 38.2 Where to download Pfam from when --bgc_bigscape_pfam is not set.
--bgc_bigscape_dir — Supply your own finished BiG-SCAPE output directory. BiG-SCAPE is then skipped.
--bgc_bigscape_cutoffs 0.3,0.5,0.7 GCF distance cutoffs to compute. They fill the dropdown in the HTML report.
--bgc_bigscape_cutoff 0.3 Which cutoff the report table shows. Must be one of the above.
--bgc_bigscape_classify none BiG-SCAPE binning mode. Only none is wired.

Pfam. BiG-SCAPE cannot run without it. If you already have a Pfam-A.hmm, pass it with --bgc_bigscape_pfam and make sure the four pressed files (.h3f .h3i .h3m .h3p) sit beside it — run hmmpress Pfam-A.hmm once if they do not.

If you do not pass it, the pipeline downloads and presses Pfam for you. That is about 400 MB to download and roughly 4 GB of free disk once unpacked and pressed. Add --save_db to keep it under <outdir>/databases/pfam, otherwise it stays in the work directory and is downloaded again on a fresh run.

The Pfam release is pinned on purpose. Pfam version changes which domains are found, which changes BiG-SCAPE distances and can change the families, so pinning keeps results reproducible between runs.

Example:

nextflow run . \
  -profile docker \
  --input example_test_data/samplesheet.csv \
  --outdir results \
  --bgc_quast_mode compare-samples \
  --run_bigscape \
  --bgc_bigscape_pfam /ABS/PATH/db/pfam/Pfam-A.hmm \
  --bgc_antismash_db /ABS/PATH/db/antismash_db_v8 \
  --bgc_deepbgc_db /ABS/PATH/db/deepbgc_db \
  --max_cpus 4 --max_memory 32.GB

Other

Option Default Meaning
--bgc_mincontiglength 3000 Contigs shorter than this are filtered out before prediction.
--save_annotations false Keep the intermediate gene-annotation files.

8. Where the results go

Everything lands under the folder you gave to --outdir.

results/
├── bgc_quast/
│   ├── compare_samples/            # compare-samples mode
│   │   ├── antiSMASH/report.tsv
│   │   ├── DeepBGC/report.tsv
│   │   └── GECCO/report.tsv
│   ├── compare_tools/              # compare-tools mode
│   │   └── <sample>/report.tsv
│   ├── compare_to_reference/       # compare-to-reference mode
│   │   ├── antiSMASH/report.tsv
│   │   ├── DeepBGC/report.tsv
│   │   └── GECCO/report.tsv
│   ├── quast/                      # only in compare-to-reference mode
│   └── bigscape/                   # only with --run_bigscape; includes BiG-SCAPE's own index.html
├── databases/
│   └── pfam/                       # only with --save_db; the downloaded and pressed Pfam
└── pipeline_info/                  # run reports, timeline, and DAG diagram

Only the folder for the mode you ran will be present. The main results are the report.tsv files.

9. Resuming a run

If a run stops partway, you can continue from where it left off by adding -resume:

nextflow run . -profile docker --input ... --outdir results -resume

Keep the same --outdir. Do not change database options on a -resume.

10. Troubleshooting

  • "command not found: nextflow" — install Nextflow, then re-run export NXF_VER=25.10.5.
  • Docker errors / nothing runs — make sure Docker is installed and running before you start.
  • Wrong Nextflow version — this pipeline needs 25.10.5. Newer versions (26.04+) will fail. Always set export NXF_VER=25.10.5.
  • Database errors from antiSMASH or DeepBGC — check the paths you passed are absolute (start with /) and point to the correct folders. The antiSMASH database must be the v8 database.
  • A run failed and you want detail — add --bgc_quast_debug to see the raw error message.
  • "Pipeline did NOT complete successfully" (red box) — the run finished but produced no comparison (for example, all contigs were shorter than --bgc_mincontiglength, so nothing reached the tools). Check your input genomes.

11. Quick reference

# 1. one-time setup
git clone <YOUR_REPO_URL> bgc_quast_ppl
cd bgc_quast_ppl
# download antiSMASH v8 and DeepBGC databases (see section 3)

# 2. every run
export NXF_VER=25.10.5

nextflow run . \
  -profile docker \
  --input data/samplesheet.csv \
  --outdir results \
  --bgc_quast_mode compare-samples \
  --bgc_antismash_db /ABS/PATH/db/antismash_db_v8 \
  --bgc_deepbgc_db /ABS/PATH/db/deepbgc_db \
  --max_cpus 4 \
  --max_memory 24.GB

About

nextflow pipeline for BGC-QUAST tool

Resources

Readme
Activity

Stars

1 star

Watchers

1 watching

Forks

0 forks
Report repository

Releases

Packages

Contributors

Languages