welch-lab/liger

What it solves

LIGER addresses the challenge of integrating and analyzing multiple single-cell datasets that may differ by experimental batch, individual, sex, tissue, species, or modality (such as RNA-seq, ATAC-seq, or DNA methylation).

How it works

The project uses integrative non-negative matrix factorization (iNMF) to identify both shared and dataset-specific factors across different data sources. It includes specialized methods like Consensus iNMF for higher confidence results and centroid alignment for improved batch effect removal and biological information conservation.

Who it’s for

It is designed for researchers performing single-cell multi-omic integration and analysis to define cell types and compare features across diverse biological datasets.

Highlights

• Supports integration across different modalities (scRNAseq, spatial transcriptomics, scMethylation, scATAC-seq).
• Enables cross-species analysis (e.g., mouse and human).
• Provides tools for clustering, identifying gene markers, and visualization via t-SNE and UMAP.
• Interfaces with existing single-cell analysis packages like Seurat.