lib5c.algorithms.variance.cross_rep module

lib5c.algorithms.variance.cross_rep.cross_rep_variance(obs, model='lognorm', min_disp=1e-08, method='mme')

Estimates pixel-wise variance across replicates.

Parameters

• obs (dict or list of np.ndarray) – Dict values or list entries are are square, symmetric count matrices across replicates.

• model ({'lognorm', 'nbinom', 'norm'}) – Statistical model to use.

• min_disp (float) – Force a minimum value of the dispersion parameter.

• method ({'mme', 'mle'}) – When model=’nbinom’, pass ‘mle’ to run maximum likelihood estimation for each pixel independently. Pass ‘mme’ to use method-of-moments variance estimation. Has no effect if model=’lognorm’.

Returns

The first three elements are the mean parameter estimate, dispersion estimate, and variance estimate, respectively, for each pixel. The fourth element is a boolean matrix showing which pixels are considered to be overdispersed.

Return type

tuple of np.ndarray