Skip to content

scMultiBench

scMultiBench is a benchmark of single-cell multimodal integration methods. The multibench package runs them from Python and scores their output with the benchmark's metrics.

conda create -n multibench python=3.12 -y
conda activate multibench
pip install multibench-sc

Python 3.9 or newer. Without conda, see Installation.

Get started How it works

1 Run a method
Many methods, one API Vertical, diagonal, mosaic and cross integration through the multibench API. The methods run in separate environments.
Benchmark metrics Clustering scores (ARI, NMI, ASW, iASW, iF1, cLISI), batch-correction scores (ASW_batch, ARI_batch, NMI_batch, GC, iLISI, kBET) and classification scores (OCA, ACA, Spec, Sens, F1), combined into rank scores per method.
Compare methods Rank methods for your data with mtb.rank, or browse the full rankings in the explorer.