simulate_whx: Simulate factor matrices and data using priors

In hjunwoo/ccfindR: Cancer Clone Finder

Description

Under Bayesian formulation, use prior distributions of factor matrices and generate simulated data

Usage

simulate_whx(nrow, ncol, rank, aw = 0.1, bw = 1, ah = 0.1, bh = 1)

Arguments

nrow	Number of features (genes).
ncol	Number of cells (samples).
rank	Rank (ncol of W, nrow of H).
aw	Shape parameter of basis prior.
bw	Mean of basis prior. Scale parameter is equal to aw/bw.
ah	Shape parameter of coefficient prior.
bh	Mean of coefficient prior. Scale parameter is equal to ah/bh.

Details

Basis W and coefficient matrices H are sampled from gamma distributions (priors) with shape (aw,ah) and mean (bw,bh) parameters. Count data X are sampled from Poisson distribution with mean values given by WH.

Value

List with elements w, h, and x, each containing basis, coefficient, and count matrices.

Examples

set.seed(1)
x <- simulate_whx(nrow=50,ncol=100,rank=5)
s <- scNMFSet(count=x$x)
s <- vb_factorize(s,ranks=seq(2,8),nrun=5)
plot(s)  

hjunwoo/ccfindR documentation built on Oct. 4, 2019, 10:31 a.m.