Analysis of Single-Cell CRISPR Screen Data • sceptre

sceptre implements an analysis pipeline for single-cell CRISPR screen (i.e., Perturb-seq) data, including data import, gRNA-to-cell assignment, quality control, and testing association between CRISPR perturbations and changes in gene expression. It is built on three principles:

Statistical rigor. Association testing is built on a principled resampling-based methodology. Built-in calibration and power checks let you verify that false positives are controlled and that real effects are detectable — on your own data.
Massive scalability. Optimized C++ routines improve performance. Optional disk-backed data structures allow the analysis of datasets too large to fit in memory. The companion sceptre Nextflow pipeline distributes analyses across hundreds of processors on a cluster or cloud.
Ease of use. A small, pipe-friendly set of functions takes you from raw count matrices to results in a handful of steps.

Installation

Install the development version of sceptre from GitHub:

# install.packages("remotes")
remotes::install_github("Katsevich-Lab/sceptre")

Documentation

New to sceptre? Get started offers a brief tour of the pipeline, with plots and explanation at each step. Other resources are the function reference and the comprehensive sceptre manual.

Featured publications

Barry et al., 2024. “Robust differential expression testing…”. Genome Biology.
Barry et al., 2024. “Exponential family measurement error models…”. Biostatistics.
Morris et al., 2023. “Discovery of target genes and pathways…”. Science.
Barry et al., 2021. “SCEPTRE improves calibration and sensitivity…”. Genome Biology.

sceptre is also featured in a 10x Genomics analysis guide.

Bug reports, feature requests, and software questions

For bug reports, please open a GitHub issue. For questions about sceptre functionality, documentation, or how to apply it to your data, please start a discussion under Q&A.