8000 GitHub - Schlieplab/ASOkai · GitHub
[go: up one dir, main page]

Skip to content
SchlieplabPublic

About

Resources

Stars

3 stars

Watchers

1 watching

Forks

Repository files navigation

ASOkai

ASOkai is an open-source pipeline for antisense-therapeutic design and analysis. Its steps can be run individually, grouped into tasks, or composed into workflows. ASOkai resolves the required upstream steps, validates their configuration and file contracts, and executes the resulting plan through CWL.

Installation

Create the Conda environment so that binary dependencies such as KMC and Bowtie 2 are installed alongside the Python dependencies, then install ASOkai into it:

conda env create -f environment.yaml
conda activate ASOkai
python -m pip install -e .

Quick start

Copy the maintained example configuration and inspect the available pipeline units:

cp examples/config.example.yaml config.yaml
ASOkai list steps
ASOkai list tasks
ASOkai list analyses
ASOkai list workflows

Preview or run the standard workflow:

ASOkai run --workflow standard --dry-run
ASOkai run --workflow standard

The pipeline reference documents the built-in steps from their runtime declarations. The example configuration shows the complete built-in configuration structure.

Antisense therapeutic sequences

ASOkai represents nucleotide sequences with T, not U.

For every AntisenseTherapeutic, the sequence property returns the antisense-strand sequence used for target recognition. For a multi-stranded therapeutic such as siRNA, it does not combine the antisense and sense strands; the sense-strand sequence is available separately as sense_strand.sequence.

Candidate generation writes therapeutics and their site bindings to separate files. Each binding stores a therapeutic ID and a site ID.

The candidate-generation step currently supports unmodified DNA ASOs and wing-gap-wing gapmer ASOs. The example configuration selects an RNase-H1 gapmer:

parameters:
  antisense:
    modality:
      type: aso
      mechanism: rnase-h1
      design:
        type: gapmer
        left_wing_length: 3
        gap_length: 10
        right_wing_length: 3

Each modality groups its mechanism under modality.mechanism and its design settings under modality.design. For an unmodified DNA ASO, set modality.design to {type: unmodified}.

ASOkai CLI

Configuration

The CLI reads ./config.yaml when present, including for CLI-selected runs. Use --configfile to select another file. For a CLI-only run, use --no-config; if no config.yaml exists, CLI selections also work without that flag. Both cases use ./data as the data directory. A file containing runs cannot be combined with CLI selections. Place --param PATH=VALUE before the first selection to set a shared parameter; omit the parameters. prefix. Use the same path after a step or task to override that run's parameter:

ASOkai run --no-config --param genome.assembly=GRCh39 --param genome.release=115 \
  --param genome.species=Homo_sapiens --step create-target-gene \
  --param target.target_id=ENSG00000133703 --param target.k=20 \
  --param target.region=transcriptomic

Run ASOkai describe step NAME to see its parameters, inputs, and outputs. Add -v to see configuration keys, input override options, and output paths.

Single-node resource requests are configured under resources:

resources:
  default:
    n_workers: 1
    memory_mb: 4096
  compute-intrinsic-features:
    n_workers: 4

Missing per-step settings inherit from resources.default. If the section is omitted, ASOkai uses 4096 MiB and one worker where parallel execution is supported. A step without worker support requests one core and does not accept a step-level n_workers setting.

Commands

Use list and describe instead of relying on a hard-coded catalog; installed plugins appear in the same commands:

# List available units
ASOkai list steps
ASOkai list tasks
ASOkai list analyses
ASOkai list workflows

# Describe one unit
ASOkai describe step download-genome
ASOkai describe task instantiate-target-gene
ASOkai describe analysis compute-intrinsic-features
ASOkai describe analysis find-repeated-sites
ASOkai describe workflow standard

# Run one step, task, or workflow
ASOkai run --step download-genome
ASOkai run --task instantiate-target-gene
ASOkai run --workflow standard

# Run multiple selected units
ASOkai run --step generate-therapeutic-candidates --step compute-intrinsic-features

# Export the selected goals as a CWL job bundle without executing them
ASOkai export --workflow standard

# Include the complete producer closure in the exported bundle
ASOkai export --step compute-intrinsic-features --recursive

# Require dependencies to exist instead of scheduling missing ones
ASOkai run --step compute-intrinsic-features --non-recursive

Third-party software and licenses

External programs used by some workflows are not installed by the Python package. Obtain binaries according to your setup and respect each upstream license.

KMC

KMC is GPLv3-only; see the upstream repository for the license text.

Bowtie 2

Bowtie 2 is GPLv3 software. ASOkai invokes its command-line executables but does not distribute or link against them.

ViennaRNA

ViennaRNA computes the binding free energies used by the potential-repeated-sites search. ASOkai imports its Python bindings, which are distributed under ViennaRNA's own license; review it before use.

Contributor setup

Install ASOkai with its contributor-only tools in the Conda environment:

conda env create -f environment.yaml
conda activate ASOkai
python -m pip install -e ".[dev]"

The contributor guide explains the pipeline contracts and provides a copyable analysis-step starting point.

Documentation

Build the documentation with the active environment's Python:

python -m sphinx -W --keep-going -b html docs/source docs/build/html

License

ASOkai is licensed under LGPL-3.0-or-later.

Copyright 2026 Alexander Schliep

About

Resources

Stars

3 stars

Watchers

1 watching

Forks

Releases

Packages

Contributors

Languages

0