GitHub

Juscan.jl is a Julia implementation of Scanpy, tailored for single-cell data analysis. Currently in its development (dev) version, the library aims to deliver high-performance, flexible, and extensible tools for preprocessing, dimensionality reduction, clustering, and visualization of single-cell datasets.

Features

  • Single-Cell Data Analysis Provides functionalities for data normalization, log transformation, dimensionality reduction (e.g., PCA), and graph-based clustering (e.g., Leiden algorithm).

  • High Performance Computing Leverages Julia's computational strengths to efficiently handle large-scale single-cell datasets.

  • Modular and Extensible Designed with a modular architecture, allowing users to easily customize and extend functionality to suit diverse analytical requirements.

Development Status

Juscan.jl is under active development. Some features are still being refined and may change over time. We welcome feedback, bug reports, and contributions from the community.

Installation

Installing the Development Version via Julia Package Manager

You can install the development version using Julia's package manager:

using Pkg
Pkg.add(url="https://github.com/zehua0417/Juscan.jl")

Cloning the Repository for Local Development

git clone https://github.com/zehua0417/Juscan.jl.git
cd Juscan.jl
julia --project -e 'using Pkg; Pkg.instantiate()'

Quick Start Example

Here’s a simple example demonstrating basic single-cell data preprocessing and analysis with Juscan.jl:

using Juscan

# Load the dataset
adata = Juscan.readh5ad("data/data.h5ad")

# Quality control
Juscan.Pp.filter_cells!(adata, min_genes=100, max_genes=8000, max_counts=140000)
Juscan.Pp.filter_genes!(adata, min_cells=3)
adata.var.mt = startswith.(adata.var_names, "MT-")
Juscan.Pp.calculate_qc_metrics!(adata, qc_vars=["mt"])

# Normalization and log transformation
adata.layers["normalized"] = deepcopy(adata.X)
Juscan.Pp.normalize_total!(adata, target_sum=1000, layer="normalized")
Juscan.Tl.logp1_transform!(adata, layer="normalized", key_added="normalized_logp1")

# Highly variable genes and PCA
Juscan.Tl.highly_variable_genes!(adata, n_top_genes=2000, layer="normalized_logp1")
Juscan.Tl.pca!(adata; layer="normalized_logp1", key_added="pca", n_pcs=15)

# Clustering and UMAP visualization
Juscan.Tl.clustering!(adata, method="km", use_pca=15, cluster_K=4)
Juscan.Tl.umap!(adata; key_added="umap", n_pcs=15, use_pca="pca")
Juscan.Pl.plot_umap(adata, color_by="clusters_0.5")

Contributing

We welcome contributions of all kinds:

  • Bug Reports & Suggestions: Please open an issue on GitHub.
  • Code or Documentation Updates.
  • Discussions: Share your feedback and experiences to help improve the project.

License

Juscan.jl is licensed under the MIT License.

Contact

If you have any questions or suggestions, please open an issue on GitHub or contact us at my e-mail.

Acknowledgments

We would like to extend our sincere gratitude to the open source community for their continuous support and inspiration. In particular, we thank:

  • Scanpy for its groundbreaking approach to single-cell analysis,
  • AnnData for providing a robust data structure for annotated data,
  • Muon.jl for its innovative multi-omic analysis tools, and
  • Automatic Single-cell Toolbox (ASCT)：An automated single-cell data analysis toolbox inspired by Seurat v4 in R.
  • scVI.jl for offering valuable insights into variational inference methods.

Their contributions and ideas have been instrumental in shaping the development and direction of Juscan.jl.

Thank you for your interest in Juscan.jl. We hope it will empower your single-cell data analysis projects!