liver-annotation 0.1.2

A machine learning model for classification of cells and annotation of clusters in scRNA-seq data from liver samples.

Liver Annotation

Liver Annotation is a Python package designed to annotate clusters in single-cell RNA sequencing (scRNA-seq) data from liver samples. This package provides a machine learning model that is specifically trained on liver cells, enabling out-of-the-box functionality without the need for pre-existing expert-annotated data.

Features

• Machine learning model trained specifically on liver cells.
• Supports both neural network and random forest classifier models.
• Annotates clusters using either the most common annotation or probability-based methods.

Installation

To install the package, use pip:

pip install liver_annotation

Usage

Classification of Cells

You can classify cells by cell type using the classify_cells function. The function requires an input in_data which is a standard scanpy/anndata object with gene expression data.

from liver_annotation import classify_cells

# Example usage
classify_cells(ann_data_obj, species="human", model_type="nn")

• species: Choose between "human" or "mouse".
• model_type: Choose between "rfc" (random forest classifier) or "nn" (neural network).

Cluster Annotation

Annotate clusters using the cluster_annotations function. This function requires an input in_data and allows you to specify the clustering algorithm and model type.

from liver_annotation import cluster_annotations

# Example usage
cluster_annotations(in_data, species="human", clusters="louvain", algorithm="mode", model_type="nn")

• clusters: The column in in_data.obs to use for cluster data.
• algorithm: Choose between "mode" or "prob" for cluster annotation.
• model_type: Choose between "rfc" or "nn".

Dependencies

• torch
• joblib
• scipy
• numpy
• scanpy

License

This project is licensed under the MIT License - see the LICENSE file for details.

Contributing

Contributions are welcome! Please open an issue or submit a pull request for any improvements or bug fixes.

Contact

For any questions or issues, please contact Madhavendra Thakur at [e-mail].