omics-multiome

Integrate paired single-cell RNA and ATAC-seq data into a unified MuData object to perform joint embedding using WNN or MultiVI latent representation. This fuses RNA and chromatin accessibility signals for unified multi-modal analysis.

Infer gene regulatory networks from paired multiome data by supporting SCENIC+ pipelines for enhancer-driven gene regulatory network (eGRN) discovery. This links chromatin accessibility peaks to gene expression within the same nuclei.

You need a pinned scverse environment and external pipeline support to run multiome analysis with MultiVI and WNN clustering. Heavy computational tasks like motif enrichment and topic modeling require these external dependencies.

You can use h5ad files to assemble a paired RNA and ATAC dataset into a MuData object. The framework processes these input files to facilitate downstream multi-modal bioinformatics workflows.

Limitations of running SCENIC+ and topic modeling include the requirement for external pipeline support for heavy computational tasks. The skill handles joint embedding but relies on external dependencies for motif enrichment.

MultiVI provides a latent representation for single-cell multiome joint embedding, whereas WNN performs weighted fused clustering. Both methods integrate RNA and chromatin accessibility signals into a unified MuData object.