decon_devmethod_workflow

Nextflow workflow for deconvolution method development and benchmarking.

Author: Sean Maden

Acknowledgements: Hédia Tnani, Nick Eagles

Setup

Virtual environment and dependencies

Scripts provided in the sh and yml folders show which dependencies are required and how to install them. For most runs, you will at minimum need a recent version of NextFlow, R, and the nnls, SummarizedExperiment, and SingleCellExperiment libraries, with additional required dependencies for other deconvolution methods besides the non-negative least squares (NNLS) method.

You can use the provided .yml file to set up a conda virtual environment. From the top level of r-nf_deconvolution, run the following:

conda env create -f ./yml/r-nf_deconvolution.yml

Activate the new environment with:

conda activate r-nf_deconvolution

You should now be ready to run nextflowr-deconvolution.

Obtaining example data

Example datasets are contained in the data folder. Several setup .R scripts are provided in the rscript folder to download and prepare example data for a workflow run.

The lung adenocarcinoma dataset from sc_mixology (Tian et al 2019) can be downloaded and set up by running:

Rscript ./rscript/prepare_lung-adeno-example.R

Quick start

Use a terminal to navigate to the top level of nextflowr-deconvolution and run the following:

sh ./sh/r-nf.sh

This should use example data to produce a series of outputs. The main outputs are stored at the top level in a .csv file called results_table_*.csv.