gibbs.samp: gibbs.samp

In rsilvabioinfo/ProbMetab: Probabilistic annotation of LC-MS based metabolomics

Description

Call a C++ function that runs a Gibbs Sampler algorithm to sample from the distribution of metabolite to compound attribution with the previous assumption that the connected combination of attributions makes more sense.

Usage

gibbs.samp(x, y, N, w, p)

Arguments

x	a vector of masses (unique from mass/retention time pairs).
y	a vector of candidate compounds for each mass.
N	number of iterations to sample.
w	matrix of compound connections.
p	matrix of likelihood probabilities.

Value

A list of matrices with attribution indexes and probabilities.

Examples

# Not to run 
# tmp <- which(w==1, arr.ind=TRUE)
# w_sparse <- sparseMatrix(i = tmp[,1], j = tmp[,2],x = rep(1, nrow(tmp)), dims = c(C,C))
 # x <- 1:ncol(wl$wm)
 # y <- 1:nrow(wl$wm)
 # system.time(conn <- gibbs.samp(x, y, 5000 , w_sparse, p))

rsilvabioinfo/ProbMetab documentation built on May 28, 2019, 3:32 a.m.