omicverse.metabol.cv_filter

Contents

omicverse.metabol.cv_filter#

omicverse.metabol.cv_filter(adata, *, qc_mask=None, cv_threshold=0.3, across='qc')[source]#

Drop features with coefficient-of-variation above cv_threshold.

Parameters:

  • qc_mask (str | ndarray | None (default: None)) – Either the name of a boolean column in adata.obs (True = QC pool sample), or a boolean array of length adata.n_obs. Required when across='qc' (the default). Ignored when across='all'.

  • cv_threshold (float (default: 0.3)) – Features with std/mean > cv_threshold are dropped. Default 0.30 is the community standard for untargeted LC-MS with QC pools; for across='all' on biological samples a much higher threshold (1–2) is typical because biology itself adds variance.

  • across (str (default: 'qc')) –

    • "qc" (default) — compute CV across QC-pool samples only (the MetaboAnalyst convention for LC-MS); requires qc_mask.

    • "all" — compute CV across every sample. Use for NMR datasets or any workflow without pooled QC injections. MTBLS1-style studies fall in this bucket.

Returns:

Subset to features that passed. adata.var['qc_cv'] carries the CV values on the kept features.

Return type:

AnnData

Contents