tuqiang2014 / moana Goto Github PK

This repository contains a Python implementation of Moana (Multi-resolution transcriptOmic ANAlysis framework), a framework for defining and predicting cell types in single-cell RNA-Seq data (Wagner and Yanai, 2018).

Here are two demonstrations of how Moana can be used to analyze scRNA-seq data:

1. Predict cell types using a Moana PBMC classifier (PBMC-4k dataset)

2. Perform clustering for human PBMCs (PBMC-4k dataset)

Additional notebooks can be found in a separate repository.

Moana requires Python version 3.5, 3.6, or 3.7, as well as the Python packages pandas, scikit-learn, and plotly.

The easiest way to install Python as well as these packages is to download and install Anaconda. Anaconda is a distribution of Python that already includes a lot of packages, including pandas, scikit-learn, and plotly. Alternatively, you can download and install Miniconda, and use the conda command to create a new Python 3 environment and install the required packages. The latter option takes up less disk space but also requires some knowledge of how to use conda, the package/environment manager that underlies both Anaconda and Miniconda.

To install Moana, make sure you have activated/selected the correct conda environment, and then type:

$ pip install moana

This is the initial release of Moana. We used this version to construct the cell type classifiers described in our preprint. Additional documentation is forthcoming.