broadinstitute/CellBender

What it solves

CellBender eliminates technical artifacts and systematic background noise from high-throughput single-cell RNA sequencing (scRNA-seq) data. Specifically, it addresses the problem of ambient RNA molecules and random barcode swapping that can contaminate raw count matrices.

How it works

It provides a remove-background module that processes raw UMI-based scRNA-seq count matrices to remove these artifacts. The tool is designed to be run as a command-line interface (CLI) and supports GPU acceleration via CUDA to handle the computational demands of the denoising process.

Who it’s for

Researchers and bioinformaticians working with single-cell experiments, including those using scRNA-seq, snRNA-seq, and CITE-seq technologies.

Highlights

• Broad Compatibility: Works across multiple sequencing modalities including scRNA-seq, snRNA-seq, and CITE-seq.
• GPU Acceleration: Supports CUDA for faster processing of large datasets.
• Flexible Deployment: Available via pip, Docker images, and WDL workflows for platforms like Terra and Dockstore.