Skip to contents

recall (Calibrated Clustering with Artificial Variables) Reproducibility

Introduction

We hope to make it as simple as possible to reproduce the results found in our paper Artificial variables help to avoid over-clustering in single-cell RNA-sequencing. To that end, we have organized our analysis scripts as a series of R vignettes in this repository.

Reproducing Figures from the recall manuscript

Clone the repository:

git clone https://github.com/lcrawlab/recallreproducibility

Then, navigate to the repo directory and launch R

cd recallreproducibility
R

You can build the entire website using the following line of R code.

pkgdown::build_site()

Note that each Rmarkdown file is not fully run by R by default. To properly run an an Rmarkdown file run during the website building process, you need to remove

knitr::opts_chunk$set(eval = FALSE)

from the file header. We do not recommend doing this for the vignettes that actually use recall, sc-SHC, and CHOIR for clustering because they have a long runtime. Rather, the R portions of these files should be put in a script and run using Rscript.

Relevant Citations

recall is now published in AJHG, here.

A. DenAdel, M. Ramseier, A. Navia, A. Shalek, S. Raghavan, P. Winter, A. Amini, and L. Crawford. A knockoff calibration method to avoid over-clustering in single-cell RNA-sequencing. AJHG.

Questions and Feedback

For questions or concerns with recallreproducibility or the recall R package, please contact Alan DenAdel or Lorin Crawford. Any feedback on the manuscript or figure reproducibility is greatly appreciated.