skyline-plots 0.4.2a0

Genomic plotting toolkit (Manhattan, Miami, locus-view and tracks)

Getting Started with skyline

version: 0.4.2a0

The skyline package is a genomic plotting toolkit. It was initially put together to plot Manhattan plots but can be generalised to plot general genomic "track" based data. Currently you can:

1. Plot Manhattan style plots
2. Plot Miami style plots
3. Plot skyview plots (multiple Manhattans stacked and viewed from the top)
4. Locus view type plots
5. Link tracks together with different sorts of linkers
6. Add label tracks

Architecture

Tracks are first-class objects that draw into Matplotlib Axes (they are not Axes subclasses), with hierarchical Theme styles and full access to Figure / Axes / Artists. Prefer imports from the package root (from skyline import manhattan, Theme, …).

High-level vs low-level API

Skyline offers two complementary surfaces:

Layer	What it is	When to use it
High-level	Recipe functions (manhattan, miami, locus_scatter, …) that return a PlotResult	Common figures, pipelines, and automation — few lines to a finished plot
Low-level	Track / building-block classes (GenomicScatter, GeneTrack, linkers, rulers, grids, …) with draw(ax)	Custom multi-track figures and layouts recipes do not cover
Shared	Hierarchical Theme / role-leaf styles	Styling both recipes and hand-built tracks

Recipes are the stable façade for everyday use; the low-level API is for composition and control. Matplotlib remains available underneath both (e.g. result.scatter.ax).

from skyline import HumanGRCh38, manhattan

res = manhattan(gwas_df, HumanGRCh38())
res.savefig("manhattan.png")
res.scatter.ax.axhline(y=7.3, ls="--")  # Matplotlib escape hatch

Installation

Conda

conda install -c cfin -c bioconda -c conda-forge skyline-plots

Optional Ensembl REST helpers:

pip install "skyline-plots[ensembl]"
# or install ensembl-rest-client into the same environment

Pip (from PyPI)

pip install skyline-plots
pip install "skyline-plots[ensembl]"   # optional REST features

The import package is still skyline (from skyline import …).

From a clone (development)

git clone [e-mail]:cfinan/skyline.git
cd skyline
python -m pip install -e ".[dev]"

For Ensembl-backed locus features during development:

python -m pip install -e ".[dev,ensembl]"

Example / demo data

Large example GWAS tables are not shipped in the installed package (they remain in the git tree under skyline/example_data/ for local notebooks and scripts). A small assembly coordinate file required by HumanGRCh37 / HumanGRCh38 / HumanNCBI36 is installed.

A download helper or documented data archive for demos is planned; until then use a git checkout for full examples.

Documentation

Online: https://cfinan.gitlab.io/skyline/

The docs home page mirrors this README. A progressive tutorial lives under resources/examples/ and the Sphinx “Tutorial” section. Notebook titles are tagged [high-level], [low-level], or [shared] so the layer is obvious in the sidebar:

1. High-level recipes (Manhattan, Miami, locus view)
2. Shared styles (Theme)
3. Low-level tracks (scatter, genes, skyview / linkers / grids)
4. Composition (recipe grids, hand-built multi-track figures, dual-locus capstone)

Local docs / CI dry-run (docker + pytest + Sphinx, HTML rsynced to pCloud):

cd resources/ci_cd
./build.sh

Tests

pytest tests/

Licence

GPLv3+ — see LICENCE.txt.