get.Zt: Get the matrix of subclones Z for fixed values of the weights...

In pneuvial/c3co: Cancer Cell Clonality using Copy-Number

Description

Optimisation is done by solving a L1-penalized least-square problem thanks to the glmnet package

Usage

get.Zt(Y, lambda, W, WtWm1)

Arguments

Y	a matrix with n rows (number of samples) and J columns (number of segments)
lambda	a positive scalar tuning the penalty level for fusion
W	a matrix with n rows (number of samples) and K columns (number of subclones)
WtWm1	a K x K square matrix (with K the number of subclones), precomputed to save time

Value

The transpose of Z, where Z a matrix with K rows (number of subclones) and J columns (number of segments)

Examples

J <- 11L  ## Number of segments
K <- 4L   ## Number of subclones
n <- 4L   ## Number of samples

Z <- matrix(1, nrow = K, ncol = J)
Z[2,    2] <- 2
Z[3,  5:6] <- 2
Z[4, 9:10] <- 2

W <- matrix(0, nrow = n, ncol = K)
W[1,1] <- W[2,2] <- W[3,3] <- W[4,4] <- 1
WtWm1 <- tcrossprod(backsolve(qr.R(qr(W)), x=diag(K)))

E <- matrix(rnorm(n*J, sd = 0.1), nrow = n, ncol = J)

Y <- W %*% Z + E

Zthat <- c3co:::get.Zt(Y, lambda = 0.01, W = W, WtWm1 = WtWm1)

pneuvial/c3co documentation built on May 25, 2019, 10:21 a.m.