Downstream analysis

scCS downstream analysis begins only after the supervised topology, ordering, velocity graph, anchor diagnostics, and future-fate sensitivity checks are acceptable.

Expression on the star

plot_gene_expression_star places measured expression on the same standardized cell positions used for scCS metrics. It is a visualization of expression, not a new scoring model. Use shared scales when comparing conditions and separate scales when genes have very different dynamic ranges.

Candidate commitment-associated genes

get_commitment_associated_genes estimates partial rank associations between gene expression and a fitted scCS outcome.

  • inference_unit="cell_exploratory" is candidate generation only.

  • inference_unit="replicate" aggregates biological replicate units before formal inference and is preferred when genuine replicate metadata exist.

  • Association is not proof of a causal lineage driver.

Fate markers

get_fate_markers compares each annotated terminal population with the root. These markers describe cell identity and are distinct from genes associated with variation in future-fate commitment within the root.

Enrichment

run_commitment_enrichment accepts a local mapping or GMT file. Local gene sets and an assay-specific background are preferred for publication-grade reproducibility. Remote Enrichr libraries are optional and can change over time.

Interpretation order

A defensible downstream analysis reports:

  1. velocity model and selected topology;

  2. ordering and its validation;

  3. DFFP parameters and anchor diagnostics;

  4. CFA, DFR, FFS, RC, UFP, and SOF;

  5. candidate genes or markers;

  6. gene-set source, background, and enrichment thresholds;

  7. biological replicate unit for any inferential claim.

See the pancreas and Schwann downstream-analysis tutorials for complete examples.