quantnado 0.4.0

Dataset generation and peak calling for multi-modal Next-Generation Sequencing data

QuantNado

QuantNado provides efficient Zarr-backed storage and analysis of genomic signal from BAM and bigWig files, with support for signal reduction, feature counting, dimensionality reduction, and quantile-based peak calling.

---

Installation

pip install quantnado

Requires Python 3.12 or 3.13.

---

Quick Start

Create a dataset from BAM files

from quantnado import QuantNado

qn = QuantNado.from_bam_files(
    bam_files=["sample1.bam", "sample2.bam", "sample3.bam"],
    store_path="dataset.zarr",
    metadata="samples.csv",  # optional
)

Load and analyse an existing dataset

from quantnado import QuantNado

qn = QuantNado.open("dataset.zarr")

# Aggregate signal over genomic ranges
promoter_signal = qn.reduce("promoters.bed", reduction="mean")
print(promoter_signal["mean"].shape)  # (n_promoters, n_samples)

# PCA on reduced signal
pca_obj, transformed = qn.pca(promoter_signal["mean"], n_components=10)
print(transformed.shape)  # (n_samples, 10)

# Generate a count matrix for DESeq2
counts, features = qn.feature_counts("genes.gtf", feature_type="gene")
counts.to_csv("counts.csv")

# Extract signal over a specific region
region = qn.extract_region("chr1:1000-5000")
print(region.shape)  # (n_samples, 4000)

---

Command-line Interface

QuantNado installs a quantnado command with two subcommands.

create-dataset — build a Zarr dataset from BAM files

quantnado create-dataset sample1.bam sample2.bam sample3.bam \
  --output dataset.zarr \
  --chromsizes hg38.chrom.sizes \
  --metadata samples.csv \
  --max-workers 8

call-peaks — call quantile-based peaks from bigWig files

quantnado call-peaks \
  --bigwig-dir path/to/bigwigs/ \
  --output-dir peaks/ \
  --chromsizes hg38.chrom.sizes \
  --quantile 0.98

Run quantnado --help or quantnado <subcommand> --help for full option listings.

---

API Reference

Full documentation is available at milne-group.github.io/QuantNado.

QuantNado

Method / Property	Description
QuantNado.from_bam_files(bam_files, store_path, ...)	Create a new dataset from BAM files
QuantNado.open(store_path, read_only=True)	Open an existing dataset
.reduce(ranges, reduction="mean")	Aggregate signal over genomic ranges (BED)
.feature_counts(gtf_file, feature_type="gene")	Generate a DESeq2-compatible count matrix
.pca(data, n_components=10)	Run PCA on a signal matrix
.extract_region(region)	Extract raw signal for a genomic region
.to_xarray(chromosomes)	Load dataset as lazy xarray DataArrays
.samples	List of sample names
.metadata	Sample metadata (DataFrame)
.chromosomes	Available chromosome names
.chromsizes	Chromosome sizes (dict)
.store_path	Path to the underlying Zarr store

---

Requirements

Dependency	Purpose
zarr, icechunk	Zarr v3 storage backend
xarray, dask	Lazy array operations
pandas, numpy	Data structures
pysam, bamnado	BAM file I/O
pyBigWig	bigWig I/O
pyranges	Genomic range operations
scikit-learn (via dask-ml)	PCA
typer, loguru	CLI and logging
ipykernel, jupyterlab, matplotlib	Example notebook (pip install "quantnado[example]")

---

License

MIT — see LICENSE.