omicverse.metabol.read_metaboanalyst

omicverse.metabol.read_metaboanalyst#

omicverse.metabol.read_metaboanalyst(path, *, group_col, sample_col=None, transpose=False)[source]#

Load a MetaboAnalyst-format CSV into AnnData.

MetaboAnalyst’s canonical bulk-upload layout is:

Patient ID,Muscle loss,Metabolite1,Metabolite2,...
PIF_1,cachexic,40.85,62.23,...
PIF_2,control,29.46,58.13,...

— samples in rows, one factor (group) column, rest are metabolite concentrations. group_col is required (no default) because every MetaboAnalyst dataset uses a different column name ("Muscle loss" for the Eisner-2010 cachexia demo, "Label" for most others). Pass the exact column header from your CSV.

Parameters:

  • path (str | Path) – CSV path or URL.

  • group_col (str) – Name of the factor column holding case/control labels. Required; copied to adata.obs['group'].

  • sample_col (Optional[str] (default: None)) – Name of the sample-ID column. If None, the first column is used (matches the MetaboAnalyst default).

  • transpose (bool (default: False)) – Set to True if your CSV has samples in columns and metabolites in rows — we’ll transpose before wrapping.

Returns:

adata.X is n_samples × n_metabolites floats; obs carries the group label; var is empty initially.

Return type:

AnnData

Contents