single-cell-multiomics-integration

Joint factor analysis of paired RNA and ATAC single-cell multi-omics data is performed using MOFA. This enables combined interpretation of diverse modalities from the same cells by loading datasets and selecting the paired integration workflow.

Unpaired single-cell data alignment across different experiments is achieved using GLUE. It aligns separate modalities before joint analysis, allowing you to integrate datasets that lack direct cell-level pairing.

Batch correction across multiple single-cell studies is supported using SIMBA. It harmonizes multi-batch single-modality data to remove technical variation while preserving biological signals across different experimental batches.

Cell label transfer from a reference to new single-cell datasets is executed using TOSICA. It transfers annotations to new datasets by leveraging the integrative analysis workflow to map labels across samples.

Trajectory inference with velocity data is supported via StaVIA and VIA. It implements pseudotime and lineage analysis for single-cell datasets, enabling trajectory inference directly within the multi-omics integration workflow.

PyTorch is required as a dependency to run single-cell multi-omics integration. The workflow relies on libraries like scglue, simba, and PyTorch to perform alignment, batch correction, and joint analysis.