pylimma.fitmixture

pylimma.fitmixture(log2e, mixprop, niter=4, trace=False)

Fit a mixture model by non-linear least squares.

Port of limma/R/fitmixture.R.

Parameters:

• log2e (ndarray) – log2 expression matrix, shape (nprobes, narrays).

• mixprop (array-like of float, length narrays) – Mixture proportion of sample 1 in each array.

• niter (int, default 4) – Non-linear least squares iterations.

• trace (bool, default False) – If True, print stdev / beta summaries per iteration.

Return type:

dict with keys A, M, stdev.