sc_reconstruction.metrics.load_cytokine_dict_from_csv#

sc_reconstruction.metrics.load_cytokine_dict_from_csv(path, *, celltype=None, fdr_threshold=0.05, min_genes=5)[source]#

Build a cytokine -> gene-list mapping from the Immune Dictionary CSV.

Cytokine signatures are taken from the Immune Dictionary of [Cui et al., 2024]. The CSV at analysis/data/frozen/cytokine_act_merged.csv has columns Celltype_Str, Cytokine_Str, Gene, FDR, Avg_log2FC (plus a few others).

Parameters:

  • celltype (str | None) – If given, restrict to one cell-type (e.g. "B_cell"). If None, cytokine signatures are pooled across all cell types.

  • fdr_threshold (float | None) – Drop genes with FDR > fdr_threshold before grouping. Set to None to keep all genes.

  • min_genes (int) – Drop cytokines with fewer than this many surviving genes.

  • path (str | Path)

Return type:

dict[str, list[str]]