Pipelines

Every analysis panel on the study page runs one of these pipelines. This reference lists what each one produces.

Pipeline index

Pipeline	Outputs
`immune_composition`	Population counts, proportions, group summaries
`immune_annotation`	Reference-atlas labels, probabilities, confidence categories, marker evidence, compatibility warnings, hierarchy labels, state programs, cell preview
`repertoire_analysis`	Clonotypes, diversity, subsampled rarefaction, CDR3 length/property/k-mer summaries, positional entropy, V/J pairing, sample distances, expansion, overlap, barcode clone maps, clone occupancy, clonal bias, clone-flow/network edges, VDJ ambiguity flags
`cell_communication`	Complex-aware ligand-receptor interactions, pathways, cytokine networks, p/q-values, specificity, role summaries, deltas
`ligand_activity`	Ligand activity ranking (AUPR-corrected / Pearson + permutation p/q), receiver target-gene support, ligand-target matrix rows, ligand-receptor-target paths, and prioritized sender→receiver pairs
`disease_workflows`	Trajectory, spatial niche, graph neighborhoods, neighborhood enrichment, TLS candidates, cytokine/state gradients, region summaries, proximity, cohort summaries
`multimodal_profiling`	ADT, paired RNA/protein CITE-seq summaries, background correction, RNA/protein discordance, phenotype panels, scATAC QC/LSI/marker peaks/gene activity/peak-to-gene/group coverage summaries, multiome, FlowSOM-style cytometry clusters/metaclusters/MST/abundance summaries
`interpret`	Grounded interpretation with citations and audit artifacts
`report`	Report, methods text, tables, figure specs, provenance

Every run records platform_version and the submitted parameter JSON. Use run IDs, pipeline names, versions, and artifact paths in methods sections.

Immunology is the canonical owner for:

Oncology imports these shared engines instead of duplicating the logic.