concord: concordance analysis

In mengchen18/omic3plus: Different methods for the integrative analysis multiple omics data

Description

concordance analysis

Usage

concord(x, y, ncomp = 2, dmod = 1, center.x = TRUE, scale.x = FALSE,
  center.y = TRUE, scale.y = FALSE, option = "uniform", kx = "all",
  ky = "all", wx = 1, wy = 1, pos = FALSE, verbose = TRUE,
  init = c("svd", "average")[2], maxiter = 1000, ncores = 1, fold = 5,
  nstart = 1, seed = NULL, loorss = FALSE, scan = TRUE, nsd = 1)

Arguments

x	a list of predictive matrices. The columns are observations, rows are varaibles. The columns (observations) has to be matched.
y	a response matrix. Rows are variables, columns are observations. The columns should be matched with columns in x.
ncomp	the number of components want to retain
dmod	the deflation mode, dmod = 2 is the original publication of
center.x	logical values, whether the variables in x should be centered
scale.x	logical values, whether the variables in x should be scaled
center.y	logical values, whether the variables in y should be centered
scale.y	logical values, whether the variables in y should be scaled
option	the option for normalizing matrix
kx	the number (if it is an integer > 1) or the proportion (if 0 < ky < 1) of kept variables in x. It should be a numeric value.
ky	the number (if it is an integer > 1) or the proportion (if 0 < ky < 1) of kept variables in y. It should be a numeric value.
wx	weight for the rows of x
wy	weight for the rows of y
pos	logical value, whether only non-negative loadings retained
verbose	if the process of calculation should be printed
init	how to initialize the algorithm. if no sparsity, svd is fast.
maxiter	maximum number of iterations allowed
ncores	number of cores to be used, passed to mclapply
fold	the number of fold to be used in cross-validation, only used if kx or ky is a vector
nstart	how many time the k-fold cross validation should be done
seed	set seed for random number generation
loorss	if the Leave-one-out procedure should be used in matrix reconstruction
scan	If the PRESS plot should be shown and used to determine the optimal k in CV
nsd	the the n*sd for selecting k automatically

Author(s)

Chen Meng

Examples

library(omic3plus)
data("NCI60_4arrays")

y <- as.matrix(NCI60_4arrays$agilent)
x <- lapply(NCI60_4arrays[2:4], as.matrix)

# no sparsity
con1 <- concord(x, y, ncomp = 3)

# sparsity on rows of x, select 10% genes
con <- concord(x, y, ncomp = 3, kx = 0.1)

# sparsity on rows of both x and y, select 10% genes
con <- concord(x, y, ncomp = 3, kx = 0.1, ky = 0.1, option = "nk")
plot(con$score.x[, 1], con$score.y[, 1])
abline(a = 0, b = 1)

mengchen18/omic3plus documentation built on May 6, 2019, 4:59 p.m.