kohlkopf / scrublet

Detect doublets in single-cell RNA-seq data

Scrublet

Single-Cell Remover of Doublets

Python code for identifying doublets in single-cell RNA-seq data. For details and validation of the method, see our preprint on bioRxiv.

Quick start:

For a typical workflow, including interpretation of predicted doublet scores, see the example notebook.

Given a raw (unnormalized) UMI counts matrix counts_matrix with cells as rows and genes as columns, calculate a doublet score for each cell:

import scrublet as scr
scrub = scr.Scrublet(counts_matrix)
doublet_scores, predicted_doublets = scrub.scrub_doublets()

scr.scrub_doublets() simulates doublets from the observed data and uses a k-nearest-neighbor classifier to calculate a continuous doublet_score (between 0 and 1) for each transcriptome. The score is automatically thresholded to generate predicted_doublets, a boolean array that is True for predicted doublets and False otherwise.

Installation:

git clone https://github.com/AllonKleinLab/scrublet.git
cd scrublet
pip install -r requirements.txt
pip install --upgrade .

Old versions:

Previous versions can be found here.

Other doublet detection tools:

DoubletFinder
DoubletDecon
DoubletDetection