https://github.com/angelovangel/nxf-alignment
Nextflow pipeline to process ONT adaptive sampling runs (basecalling, mapping, statistics)
https://github.com/angelovangel/nxf-alignment
basecalling nanopore nextflow-pipeline sequencing
Last synced: 3 months ago
JSON representation
Nextflow pipeline to process ONT adaptive sampling runs (basecalling, mapping, statistics)
- Host: GitHub
- URL: https://github.com/angelovangel/nxf-alignment
- Owner: angelovangel
- License: mit
- Created: 2025-12-02T06:42:50.000Z (6 months ago)
- Default Branch: main
- Last Pushed: 2026-01-13T13:00:33.000Z (5 months ago)
- Last Synced: 2026-01-13T21:22:40.748Z (5 months ago)
- Topics: basecalling, nanopore, nextflow-pipeline, sequencing
- Language: Python
- Homepage:
- Size: 174 KB
- Stars: 0
- Watchers: 0
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# nxf-alignment
A Nextflow workflow for basecalling (ONT only), aligning, and variant calling for long-read sequencing data (ONT + HiFi).
## Features
- **Basecalling**: Uses Dorado for basecalling (and demultiplexing) with optional adaptive sampling support (ONT)
- **Alignment**: Aligns reads to a reference genome using Dorado aligner (modifications are preserved, ONT or HiFi data)
- **Coverage Analysis**: Calculates per-region coverage statistics with thresholds (1x, 10x, 20x, 30x)
- **SNP Variant Calling**: Uses Clair3 or DeepVariant for SNP variant calling (ONT or HiFi data)
- **Structural Variant Calling**: Uses Sniffles2 for structural variant calling (ONT or HiFi data)
- **Variant Annotation**: Uses snpEff for variant annotation (ONT or HiFi data)
- **Base Modifications Analysis**: Uses modkit for base modifications analysis (ONT or HiFi data)
- **Interactive HTML Report**: Generates an interactive report with read statistics, coverage, variants and annotations metrics
>Note: You can also import this workflow in EPI2ME, see [EPI2ME documentation](https://epi2me.nanoporetech.com/)
## Requirements
- **Nextflow** >= 23.04
- **Docker**
- **NVIDIA GPU** (for basecalling and variants)
## Quick Start
#### Basic Workflow (Basecalling + Alignment + Variant Calling)
For an adaptive sampling run, basecalling is done for the accepted reads based on the decision file produced by MinKNOW.
```bash
nextflow run angelovangel/nxf-alignment \
--pod5 /path/to/pod5/dir \
--asfile /path/to/AS_decisions.csv \ # optional
--model hac \
--bed /path/to/regions.bed \ # optional, if provided the report contains coverage analysis per region from bed file
--ref /path/to/ref.fasta \
--snp \
--annotate
```
#### Barcoded run (Basecalling + Alignment)
For a barcoded run, provide a [samplesheet](#sample-sheet-barcoded-runs) and kit name
```bash
nextflow run angelovangel/nxf-alignment \
--pod5 /path/to/pod5/dir \
--model hac,5mC_5hmC \
--bed /path/to/regions.bed \
--ref /path/to/ref.fasta \
--kit SQK-RBK114-96 \
--samplesheet /path/to/samplesheet.csv
```
>Note: Sample name is obtained from the pod5 file (the sample ID entered in MinKNOW). If another sample name is desired, use the `--samplename` parameter. For barcoded runs sample names are taken from the samplesheet.
#### Skip Basecalling (Align Existing BAM/FASTQ + SNP/SV Variant Calling)
If the basecalling has been performed before, the pipeline can be run with the `--reads` parameter. The reads can be in any HTS format, a directory of reads can also be given. If the reads contain base modifications, you can use the `--mods` parameter to perform base modification analysis - create a summary of counts of modified and unmodified bases.
```bash
nextflow run angelovangel/nxf-alignment \
--reads /path/to/reads.bam \ # can be also a directory with reads
--ref /path/to/ref.fasta \
--bedfile /path/to/regions.bed \
--snp \
--sv \
--annotate \
--mods
```
#### Skip alignment (basecalling only)
Basecalling (for single sample and barcoded runs) can also be performed without alignment, using the `--basecall` or `--report` parameters.
`--basecall` will do basecalling (and evt demultiplexing), `--report` will do basecalling + report
```bash
nextflow run angelovangel/nxf-alignment \
--pod5 /path/to/pod5/dir \
--model hac \
--kit SQK-RBK114-96 \ # for barcoded runs only
--samplesheet /path/to/samplesheet.csv \ # for barcoded runs only
--report
```
## Parameters
#### Core Parameters
| Parameter | Type | Default | Description |
|-----------|------|---------|-------------|
| `pod5` | path | - | Directory containing POD5 files (required if not using `--reads`) |
| `reads` | path | null | Path to input BAM/FASTQ file(s) or directory (skips basecalling) |
| `ref` | path | - | Reference genome in FASTA format (required unless `--basecall` or `--report` is used) |
| `basecall` | boolean | `false` | Run the pipeline up to basecalling only |
| `report` | boolean | `false` | Run the pipeline up to reporting only (skips alignment and variants) |
| `model` | string | `fast` | Dorado basecall model, see [available models](https://software-docs.nanoporetech.com/dorado/latest/models/list/)|
| `outdir` | string | `results` | Output directory for results |
#### Optional (advanced) Parameters
| Parameter | Type | Default | Description |
|-----------|------|---------|-------------|
| `samplename` | string | null | Sample name to use for non-barcoded runs (if not provided, sample name is obtained from the pod5 file) |
| `asfile` | path | null | Adaptive sampling decisions CSV (if using AS filtering) |
| `herro` | boolean | null | Enable HERRO read correction. The corrected reads will be in 00-basecall, but will NOT be used in alignment. |
| `kit` | string | null | Barcoding kit name (e.g., `SQK-NBD111-96`). Required for barcoded runs |
| `samplesheet` | path | null | Sample sheet CSV or XLSX with columns: `sample`, `barcode`. Required for barcoded runs |
| `bed` | path | null | BED file with target regions (auto-generated from reference if not provided) |
| `snp` | boolean | false | Perform SNP variant calling using Clair3 or DeepVariant |
| `snp_caller` | string | `clair3` | SNP variant caller to use (`clair3` or `deepvariant`, use only with `--snp`) |
| `clair3_platform` | string | `ont` | Platform to use for Clair3 (`ont` or `hifi`) |
| `clair3_model` | string | `r1041_e82_400bps_hac_v500` | Model to use for Clair3 |
| `deepvariant_model` | string | `ONT_R104` | Model to use for DeepVariant |
| `sv` | boolean | false | Perform SV variant calling using Sniffles2 |
| `phase` | boolean | false | Perform SNP phasing using Whatshap (use only with `--snp`, only diploid cases supported) |
| `annotate` | boolean | false | Annotate SNP variants using snpEff (use only with `--snp`) |
| `anno_db` | string | `hg38` | Database to use for annotation |
| `anno_filterQ` | int | `20` | Filter out SNP variants with quality lower than this before annotation |
| `mods` | boolean | false | Perform base modification analysis using modkit (`--ref` is required)|
| `mods_filter` | int | `5` | Minimum coverage for base modifications calls |
#### Profiles
Predefined set of parameters for common use cases, use with `-profile`:
| Profile | Description |
|---------|-------------|
| `standard` | Standard workflow with Docker GPU support |
| `dev` | Workflow for testing on Apple Silicon |
| `singularity` | Workflow with Singularity GPU support |
| `revio` | Workflow optimized for HiFi Revio (use with `--reads` to skip basecalling)|
## Output Structure
```
output/
├── 00-basecall/
│ ├── reads.bam # Basecalled reads
│ ├── reads.bam.bai # BAM index
│ └── processed/ # Per-sample BAMs (if barcoded)
│ ├── sample_1.bam
│ └── sample_1.bam.bai
├── 01-align/
│ ├── reads.align.bam # Aligned reads
│ └── reads.align.bam.bai # BAM index
│ ├── reads.align.ht.bam # Haplotagged aligned reads
│ └── reads.align.ht.bam.bai # Haplotagged aligned BAM index
├── 02-coverage/
│ ├── reads.hist.tsv # Coverage histogram
│ └── reads.bigwig # Coverage bigwig
├── 03-variants/
│ ├── reads.snp.vcf # SNP variants
│ ├── reads.sv.vcf # SV variants
│ └── reads.ann.vcf # Annotated variants
├── 04-modifications/
│ ├── reads.bedmethyl # Output of modkit pileup
│ └── reads.summary.tsv # Base modification summary
├── nxf-alignment-report.html # Workflow report
├── nxf-alignment-execution-summary.txt # Workflow execution summary
└── variants-annotation-report.html # Variants annotation report
```
## Input Files
#### BED File (Optional)
Tab-separated file defining target regions:
```
chr1 1000 5000 GENE_A
chr1 8000 12000 GENE_B
```
If not provided, the workflow auto-generates a BED file covering the entire reference.
#### Sample Sheet (Barcoded Runs)
CSV with minimum columns `sample` and `barcode`:
```
sample,barcode
sample_1,barcode01
sample_2,barcode02
```
#### Adaptive Sampling Decisions File (Optional)
This file is generated by MinKNOW during an adaptive sampling run, and can be found under `runfolder/adaptive_sampling/AS_decisions.csv`
## Enabling docker GPU Support
If you observe the error "could not select device driver with capabilities gpu", additional docker setup is required. The `nvidia-container-toolkit` has to be installed and running on your system. See [here](https://epi2me.nanoporetech.com/epi2me-docs/installation/) and [here](https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/latest/install-guide.html) for details.
## Citation
If you use this workflow, please cite:
- **Dorado**: https://github.com/nanoporetech/dorado
- **Bedtools**: Quinlan & Hall, 2010
- **Nextflow**: Di Tommaso et al., 2017