checkatlas 0.0.15

One liner tool to check the quality of your single-cell atlases

CheckAtlas

CheckAtlas is a one liner tool to check the quality of your single-cell atlases. For every atlas, it produces the quality control tables and figures which can be then processed by multiqc. CheckAtlas is able to load Scanpy, Seurat, and CellRanger files.

Summary

1. Parse Scanpy, Seurat and CellRanger objects

   CheckAtlas should be able to load : .rds, .h5 and .h5ad corresponding to single-cell experiment. Need to implement :

   • automatic conversion of Seurat object to Scanpy with SeuratDisk
   • Rapid check-up of files to see if a Seurat or Scanpy can be found
   • Automatic search in Scanpy files of key information = raw data, normalized data, integrated data, reductions, layers, assays, metadatas, etc...

2. Create checkatlas summary files

   Go through all Scanpy files and extract summary information. We won't to extract :

   • All basic QC (nRNA, nFeature, ratio_mito)
   • General information (nbcells, nbgenes, nblayers)
   • All elements in scanpy objects (obs, obsm, uns, var, varm)
   • Reductions (pca, umap, tsne)
   • All metrics (clustering, annotation, dimreduction, specificity)

3. Parse checkatlas files in MultiQC

   Update MultiQC project to add checkatlas parsing. Dev project in: https://github.com/becavin-lab/MultiQC/tree/checkatlas

https://checkatlas.readthedocs.io/en/stable/

Examples

1. Evaluate and compare different atlases: https://github.com/becavin-lab/checkatlas/blob/3a4f88e94716c09a3b9c86010f570743a5855461/examples/Atlas_comparison.ipynb

https://checkatlas.readthedocs.io/en/stable/CheckAtlas_example_1/CheckAtlas_example_1.html

2. Evaluate different version of your atlas: https://github.com/becavin-lab/checkatlas/blob/3a4f88e94716c09a3b9c86010f570743a5855461/examples/Version_comparison.ipynb

https://checkatlas.readthedocs.io/en/stable/CheckAtlas_example_2/CheckAtlas_example_2.html

3. Explore Scanpy, Seurat and CellRanger objects in your folder: https://github.com/becavin-lab/checkatlas/blob/main/examples/AtlasType_comparison.ipynb

https://checkatlas.readthedocs.io/en/stable/CheckAtlas_example_3/CheckAtlas_example_3.html

Installation

CheckAtlas can be downloaded from PyPI. However, the project is in an early development phase. We strongly recommend to use the developmental version.

Install checkatlas development version

git clone [e-mail]:becavin-lab/checkatlas.git
pip install checkatlas/.

Install MultiQC with checkatlas file management. This version of MultiQC is available at checkatlas branch of github.com:becavin-lab/MultiQC.

git clone [e-mail]:becavin-lab/MultiQC.git
cd MultiQC/
git checkout checkatlas
pip install .

Install it from PyPI

pip install checkatlas

Usage

The one liner way to run checkatlas is the following:

$ cd your_search_folder/
$ python -m checkatlas .
#or
$ checkatlas .

Or run it inside your python workflow.

from checkatlas import checkatlas
checkatlas.run(path, atlas_list, multithread, n_cpus)

Development

Read the CONTRIBUTING.md file.

Project developed thanks to the project template : (https://github.com/rochacbruno/python-project-template/)