dada2-py

A pure-Python re-implementation of DADA2 (Callahan et al., Nature Methods 2016) for exact amplicon sequence variant (ASV) inference from amplicon sequencing data.

• AnnData / pandas-friendly — drop-in for downstream microbiome analysis
• No rpy2, no R install, no Bioconductor dependency
• Same API as the R dada2 workflow (filter_and_trim → learn_errors → dada → merge_pairs → make_sequence_table → remove_bimera_denovo → assign_taxonomy)
• ASV-level identity matches R DADA2 on the canonical MiSeq SOP test data

This is a standalone mirror of the canonical implementation that lives in omicverse. All algorithmic work is developed upstream in omicverse and synced here for users who want DADA2 without the full omicverse stack.

Install

pip install pydada2

Quick-start

from pydada2 import (
    filter_and_trim, learn_errors, dada,
    merge_pairs, make_sequence_table,
    remove_bimera_denovo, assign_taxonomy,
)

# 1) Quality filter + trim
filter_and_trim(
    fwd="raw/F.fastq.gz", filt="filt/F.fastq.gz",
    rev="raw/R.fastq.gz", filt_rev="filt/R.fastq.gz",
    trunc_len=(240, 160), max_ee=(2, 2), trunc_q=2,
)

# 2) Learn the error model
errF = learn_errors("filt/F.fastq.gz")
errR = learn_errors("filt/R.fastq.gz")

# 3) Run the divisive amplicon denoising algorithm
ddF = dada("filt/F.fastq.gz", err=errF)
ddR = dada("filt/R.fastq.gz", err=errR)

# 4) Merge paired-end reads
mergers = merge_pairs(ddF, "filt/F.fastq.gz", ddR, "filt/R.fastq.gz")

# 5) Sample × sequence table
seqtab = make_sequence_table(mergers)

# 6) Chimera removal
seqtab_nochim = remove_bimera_denovo(seqtab, method="consensus")

# 7) Taxonomy assignment
taxa = assign_taxonomy(seqtab_nochim, ref_fasta="silva_nr99.fa.gz")

What's included

Relationship to R DADA2

This is a Python port of the canonical R/C++ DADA2 package (benjjneb/dada2). All algorithmic behaviour is checked against the R reference on the MiSeq SOP test data (see tests/test_r_parity.py). The R reference is invoked from /scratch/users/steorra/env/CMAP during testing.

Citation

If you use this package, please cite the original DADA2 paper:

Callahan, B.J. et al. DADA2: High-resolution sample inference from Illumina amplicon data. Nature Methods 13, 581–583 (2016).

and acknowledge omicverse / this repo for the Python port.

License

LGPL-2 — matches upstream R DADA2.