omicverse.metabol.msea_gsea

Contents

omicverse.metabol.msea_gsea#

omicverse.metabol.msea_gsea(deg, *, stat_col='stat', pathways=None, n_perm=1000, min_size=3, max_size=500, seed=0, mass_db=None)[source]#

GSEA-style ranked enrichment via gseapy.prerank.

Parameters:

  • deg (DataFrame) – Output DataFrame from differential(). Rows indexed by metabolite name; column stat_col provides the ranking metric.

  • stat_col (str (default: 'stat')) – Which column of deg to rank on. Default "stat" (signed t-statistic); "log2fc" is another common choice.

  • pathways (Optional[dict[str, list[str]]] (default: None)) – Dict mapping pathway name to list of KEGG compound IDs.

  • n_perm (int (default: 1000)) – Permutation count for the empirical null. 1000 is fine for tutorials; bump to ≥10000 for publication.

  • mass_db (Optional[DataFrame] (default: None)) – Optional pre-fetched ChEBI DataFrame from fetch_chebi_compounds() — same role as in msea_ora(). Recommended for cold-cache runs to avoid per-name PubChem REST round-trips.

Returns:

Columns: Term, NES, NOM p-val, FDR q-val, ES, Lead_genes (metabolites driving the enrichment).

Return type:

pd.DataFrame

Contents