cal_E_acc: cal_E_acc

Description Usage Arguments Value Author(s)

View source: R/Oct15_complicated_w_prior_acc_met_logit_nonlinear_h1h0_no_rna_vtilde_all_link_simple_only_last_pst.R

Description

Perform E-step (i.e. calculate the expectations of missing data) for one iteration of scCAS data in the EM algorithm.

Usage

1
cal_E_acc(rij, replacef0, phi_atac, w_acc, qi)

Arguments

rij

ratio f1/(f1+f0) for scCAS data.

replacef0

(1-rij) for scCAS data.

phi_atac

the initialization of the probability for scCAS data belonging to cluster k (i.e. z[i,k]=1).

w_acc

the initialization of the probability for scCAS data that gene g is active in cluster k.

qi

the initialization of the probability that scCAS data has high gene score when gene g is active in cell i (i.e. pi_i1).

Value

prob_z_atac

Expectation of z.

prob_z_u_atac

Expectation of z*u.

prob_z_u_ut_atac

Expectation of z*u*o.

Author(s)

Jiaxuan WANGWU


cuhklinlab/scAMACE documentation built on Dec. 19, 2021, 7:03 p.m.