Multi-modal profiling

Use this workflow when immune phenotypes are supported by protein, chromatin accessibility, cytometry, or multiome data. It brings CITE-seq RNA/protein summaries, scATAC peak-matrix summaries, FlowSOM-style cytometry clustering, and dependency-gated multimodal outputs into one review path.

Research question

How do protein markers, paired RNA/protein signals, chromatin accessibility, cytometry clusters, and optional multimodal dependencies support or challenge immune phenotype assignments?

Use case

Use this when RNA-only annotation is not enough. The workflow helps review ADT background correction, RNA/protein discordance, CD-marker phenotype panels, scATAC QC and gene activity, peak-to-gene links, cytometry metaclusters, and sample-level abundance patterns.

Suggested path

1. Register cite_seq, atac, or fcs datasets.
2. Run multimodal_profiling.
3. Review ADT heatmap values, background correction, joint RNA/protein coordinates, phenotype panels, RNA/protein discordance, ATAC QC, LSI coordinates, marker peaks, gene activity, peak-to-gene links, group coverages, motif cards, cytometry coordinates, population bars, FlowSOM-style metaclusters, marker heatmap rows, and sample abundance summaries.
4. Use AI/reporting once completed runs exist.

Dependency gates

• muon for full WNN/multiome graph workflows
• SnapATAC2 for full scATAC preprocessing and peak calling
• FlowUtils or flowutils for binary FCS parsing