LIGER

LIGER performs integration of multiple single-cell datasets using integrative non-negative matrix factorization to identify shared and dataset-specific factors for unbiased cell-type definition across scRNA-seq, snATAC-seq, single-nucleus DNA methylation, spatial transcriptomics, and cross-species comparisons.

Key Features:

• Integrative Non-Negative Matrix Factorization: Uses integrative NMF to extract shared and dataset-specific latent factors across multiple datasets.
• Multi-Modal Data Integration: Supports integration of scRNA-seq, snATAC-seq, single-nucleus DNA methylation, spatial transcriptomics, and cross-species datasets.
• Analysis Workflow Components: Implements data preprocessing and normalization, joint factorization, quantile normalization, joint clustering, and visualization.
• Performance: Analysis runtime is reported to range from about 1–4 hours depending on dataset size.

Scientific Applications:

• Single-Cell Genomics: Integrates scRNA-seq and snATAC-seq to enable comprehensive profiling of cellular identities.
• Cross-Species Analysis: Compares datasets from different species to identify conserved cell types or lineage-specific adaptations.
• Epigenomic Studies: Applies to single-nucleus DNA methylation data to investigate epigenetic regulation in cell-type differentiation.
• Spatial Transcriptomics: Integrates spatial transcriptomic data to study the spatial organization and functional states of cells within tissues.

Methodology:

Performs data preprocessing and normalization, joint factorization via integrative non-negative matrix factorization, quantile normalization, joint clustering, and visualization.