msmu._tools._dea._dea

permutation_test

permutation_test(mdata, category, control, expr=None, modality='protein', n_resamples=1000, n_jobs=1, statistic='welch', force_resample=False, fdr='empirical')

Perform a permutation test on the given MuData object. Parameters

---

mdata : md.MuData The MuData object containing the data. modality : str The modality to perform the test on. Default is 'protein' category : str The category column in the mdata.obs. control : str The control group label. expr : str | None The experimental group label. If None, all other groups are considered experimental. Default is None. n_resamples : int The number of resamples for the permutation test. Default is 1000. n_jobs : int The number of parallel jobs to run. statistic : str The statistical test to use. Options are 'welch', 'student', 'wilcoxon', or 'median_diff'. Default is 'welch'. fdr: str | bool The FDR control method to use. Options are 'empirical', 'bh', 'storey'. Default is 'empirical'. force_resample : bool If True, forces resampling even if the number of permutations exceeds the possible combinations. Default is False. Returns

---

PermutationTestResult The result of the permutation test. Raises

---

ValueError If the statistic is not one of the supported types.