Atlas feature selection benchmarking

This repository contains code for benchmarking the effect of feature selection on scRNA-seq atlas construction and use.

For more information please refer to the documentation on the wiki.

Directory structure

• analysis/ - Notebooks used to perform analysis of the results
  • R/ - R functions used in the analysis notebooks
• bin/ - Scripts used in Nextflow workflows
  • functions/ - Functions used across multiple scripts
• conf/ - Nextflow configuration files
• data/ - Data files used in the analysis. Metric score files can downloaded from figshare.
  • selected-features/ - Selected features files for each dataset. Can be downloaded from figshare.
• envs/ - conda environment YAML files
• output/ - Output from Nextflow workflows (not included in git)
• reports/ - RMarkdown files and functions for output reports generated by the pipeline
• work/ - The Nextflow working directory (not included in git)
• workflows/ - Nextflow workflow files
• LICENSE - The project license
• main.nf - Main Nextflow workflow file
• nextflow.config - Main Nextflow config file
• README.md - This README
• style_bin.sh - A script for styling the files in bin/

Data files

Data files for this project are available from figshare. The analysis reports in the analysis/ directory assume:

• Metric files are downloaded to the data/ directory
• Selected features files are downloaded to the data/selected-features/ directory