liulab-data¶

Look up a GEO study, inspect its sequencing runs, and download the raw FASTQ — in a few lines of Python. liulab-data (import name labdata) talks to NCBI for you and keeps every object lazy, so nothing hits the network until you actually ask for something.

New here? Jump to the GEO Series tutorial.

Installation¶

pip install liulab-data

Downloading FASTQ also needs two external tools, sra-tools and pigz (plus curl, for the original-format download path). Recovering FASTQ from a 10x cellranger BAM also needs 10x Genomics' bamtofastq. Install them from Bioconda/conda-forge:

conda install -c bioconda -c conda-forge sra-tools pigz curl 10x_bamtofastq

Working inside the lab's pixi project?

Every tool is already in the project environment — pixi install gives you the Python package and sra-tools/pigz/bamtofastq in one step.

Configuration¶

NCBI asks every caller to identify themselves with a contact email. Set it once and you're ready:

export NCBI_EMAIL="you@lab.org"     # required by NCBI
export NCBI_API_KEY="..."           # optional — raises your rate limit

Tip

Prefer not to use environment variables? Run labdata ncbi configure to save your email to ~/.cache/liulab-data/ once and forget about it.

Next steps¶

GEO Series tutorial — from an accession to downloaded FASTQ, step by step.
10x cellranger BAM → FASTQ — recover remap-ready FASTQ when a 10x run's .sra is broken.
API reference — every class and method in full detail.