do_impute: Sampling multiple copies of imputed values and cell factors...

In xyz111131/SIMPLEs: SIMPLEs: single-cell RNA sequencing imputation and cell clustering methods by modeling gene module variation

Description

do_impute get the posterior mean and variance for the imputed values and factors given the parameters, which can be returned from SIMPLE or SIMPLE_B.

Usage

do_impute(dat, Y, beta, lambda, sigma, mu, pi, pos_mean = NULL,
  pos_sd = NULL, celltype = NULL, mcmc = 10, burnin = 2,
  verbose = F, pg = 0.5, cutoff = 0.1)

Arguments

dat	scRNASeq data matrix. Each row is a gene, each column is a cell.
Y	Initial imputed data, output from SIMPLE_B or SIMPLE.
beta	Factor loadings, output from SIMPLE_B or SIMPLE.
lambda	Variances of factors for each cluster, output from SIMPLE_B or SIMPLE.
sigma	Variances of idiosyncratic noises for each cluster, output from SIMPLE_B or SIMPLE.
mu	Mean expression for each cluster, output from SIMPLE_B or SIMPLE.
pi	Probabilites of cells belong to each cluster, output from SIMPLE_B or SIMPLE.
pos_mean	Gene mean. If centerized each gene before estimating the parameters, provide the overall mean of gene expression removed from the data matrix, output from SIMPLE_B or SIMPLE. Default is NULL.
pos_sd	Gene standard deviation. If scaled each gene before estimating the parameters, provide the overall standard deviation of gene expression removed from the data matrix, output from SIMPLE_B or SIMPLE. Default is NULL. Default is NULL.

Value

do_impute returns a list of imputation results in the following order.

1. loglikThe log-likelihood of the imputed gene expression at each iteration.

2. imptA matrix contains the expectation of imputed expression.

3. impt_varA matrix contains the variance of each imputed entry.

4. EFPosterior means of factors

5. varFPosterior covariance matrix of factors

6. consensus_clusterScore for the clustering stability of each cell by multiple imputations.

Author(s)

Zhirui Hu, [e-mail]

Songpeng Zu, [e-mail]

xyz111131/SIMPLEs documentation built on Jan. 8, 2020, 2:48 a.m.