Benchmarking out-of-reference detection

One of the goals of reference-based single-cell RNA-seq analysis is to detect altered cell states that are not observed in the reference dataset. This repository contains code to benchmark workflows for integration and differential analysis on the task of detection of Out-of-reference (OOR) states, used in our paper on reference design for disease state identification. For code to reproduce analysis in the manuscript see the reproducibility repo.

The structure of the API was inspired by the OpenProblems task structure. This package was built using the scverse cookiecutter template.

Getting started

• Quick-start example

Installation

You need to have Python 3.8 or newer installed on your system. If you don't have Python installed, we recommend installing Miniconda.

There are several alternative options to install oor_benchmark:

1. Create a new conda environment

conda create --name oor-benchmark-env python=3.10
conda activate oor-benchmark-env

2. Install R and R dependencies

conda install conda-forge::r-base==4.0.5 bioconda::bioconductor-edger conda-forge::r-statmod

3. Install the latest development version:

pip install git+https://github.com/emdann/oor_benchmark.git@master

Citation

Dann E., Cujba A.M., Oliver A.J., Meyer K., Teichmann S.A. and Marioni J.C. Precise identification of cell states altered in disease with healthy single-cell references. Nature Genetics https://doi.org/10.1038/s41588-023-01523-7