danaSeq

Metagenomic analysis pipelines for Oxford Nanopore and Illumina sequencing data. Four independent Nextflow DSL2 pipelines cover real-time read classification, long-read and short-read assembly, and downstream MAG analysis including binning, annotation, taxonomy, and metabolic profiling. Assembly pipelines produce a FASTA + depth table that feeds directly into mag_analysis.

Architecture

danaSeq/
├── nanopore_live/          Real-time analysis during sequencing
│   │                       9 modules, 14 processes -> DuckDB
│
├── nanopore_assembly/      Long-read assembly + mapping + depth
│   │                       3 modules: preprocess, assembly, mapping
│   │                       Flye/metaMDBG/myloasm -> minimap2 -> CoverM
│   │                       Output: assembly.fasta + depths.txt + BAMs
│
├── illumina_assembly/      Multi-assembler consensus + mapping + depth
│   │                       7 modules: preprocess, error_correct, normalize,
│   │                       merge_reads, assembly, dedupe, mapping
│   │                       Output: assembly.fasta + depths.txt + BAMs
│
├── mag_analysis/           Technology-agnostic downstream analysis
│   │                       10 modules: binning, annotation, taxonomy, rrna,
│   │                       metabolism, mge, eukaryotic, gene_depths,
│   │                       phylogeny, viz
│   ├── viz/                Interactive Svelte dashboard
│   └── bin/                Pipeline scripts
│
└── tests/                  Pipeline tests

Pipelines

Pipeline	Purpose	Key tools
nanopore_live	Real-time analysis during sequencing	Kraken2, Prokka/Bakta, HMMER3, DuckDB
nanopore_assembly	Long-read assembly + mapping + depth	Flye, metaMDBG, myloasm, minimap2, CoverM
illumina_assembly	Multi-assembler consensus assembly	Tadpole, Megahit, SPAdes, metaSPAdes, BBMap
mag_analysis	Technology-agnostic downstream analysis	7-binner consensus, DAS Tool, Binette, 50+ processes

Getting Started

1. Clone the repository

git clone https://github.com/rec3141/danaSeq.git
cd danaSeq

2. Choose a runtime

Method	Best for	Setup
Conda	Local/laptop development	cd <pipeline> && ./install.sh
Docker	Reproducible runs, CI	docker pull ghcr.io/rec3141/danaseq-mag-analysis:latest
Apptainer	HPC clusters	./run-*.sh --apptainer (auto-pulls SIF)

3. Download databases

# Interactive menu (shows sizes and descriptions)
./download-databases.sh

# Human reference (~4 GB, required by both assembly pipelines)
./download-databases.sh --human

# MAG analysis databases
./download-databases.sh --genomad --checkv --checkm2 --kaiju

Quality Standards

MAGs are classified per MIMAG (Bowers et al. 2017):

Quality	Completeness	Contamination	Additional
High	>90%	<5%	23S, 16S, 5S rRNA + tRNAs
Medium	>50%	<10%	--
Low	<50%	<10%	--

Resource Requirements

Configuration	CPUs	RAM	Storage
Minimum (no Kraken)	16	32 GB	500 GB
Recommended	32	128 GB	1 TB
Full analysis	32	256 GB	2 TB

License

MIT. See LICENSE.