A set of processes/pipelines for bioinformatics based on pipen

biopipen is a collection of 170+ ready-made bioinformatics processes and pipelines, built on top of the pipen workflow framework. Each process wraps a well-known bioinformatics tool (Seurat, CNVkit, bcftools, immunarch, ...) into a reusable, composable unit with typed input/output options, environment handling, and optional interactive reports powered by pipen-board.

Use processes individually as Python classes, chain them into custom pipelines, or run any of the built-in pipelines directly from the command line.

Installation

pip install -U biopipen

Optional extras:

pip install -U "biopipen[runinfo]"     # write run information to JSON files
pip install -U "biopipen[log2file]"    # log to files

Requires Python >= 3.9. Process-level dependencies (e.g. Seurat, CNVkit) are usually installed per-process in conda environments (see the documentation).

Usage

As Python APIs

Import any process, customize it with Proc.from_proc(), and run it with pipen:

from pipen import Proc, Pipen
from biopipen.ns.bed import BedLiftOver

MyBedLiftOver = Proc.from_proc(BedLiftOver)

if __name__ == "__main__":
    Pipen().set_start(MyBedLiftOver).run()

Processes are also composable — chain multiple processes together as a workflow by declaring dependencies with requires:

import pipen
from biopipen.ns.vcf import VcfLiftOver, VcfIndex

# Index the VCF lifted over by VcfLiftOver
class MyVcfIndex(VcfIndex):
    requires = VcfLiftOver

if __name__ == "__main__":
    pipen.run(
        name="liftover_pipeline",
        starts=[VcfLiftOver],
        data=[{"invcf": "input.vcf"}],
    )

The output of VcfLiftOver is automatically routed into MyVcfIndex as its input. Any chain of requires builds a full dependency graph that pipen schedules and runs.

From the command line

biopipen registers each of its namespaces as a pipen run plugin. Every process is runnable with full argument parsing (use --help to explore options):

pipen run bed BedLiftOver --in.inbed input.bed --envs.chain hg19ToHg38.over.chain

Available namespaces and their processes:

Built-in pipelines

Documentation

Full API reference, pipeline guides, and per-process configuration details are available at https://pwwang.github.io/biopipen.