Set of tools to find coherent patterns in microarray data using a Bayesian sparse latent factor model (Duarte and Mayrink 2015 - http://link.springer.com/chapter/10.1007%2F978-3-319-12454-4_15). Considerable effort has been put into making slfm fast and memory efficient, turning it an interesting alternative to simpler methods in terms of execution time. It implements versions of the SLFM using both type of mixtures: using a degenerate distribution or a very concentrated normal distribution for the spike part of the mixture. It also implements additional functions to help pre-process the data and fit the model for a large number of arrays.
Sparse Latent Factor Model
slfm is a set of tools to find coherent patterns in microarray data using a Bayesian sparse latent factor model. Considerable effort has been put into making slfm fast and memory efficient, turning it an interesting alternative to simpler methods in terms of execution time.
adding % significant column to slfm_list output
using S/N/I notation to slfm output
alpha inference now uses just the interested part of the chain
user can now choose which type of mixture to use (degenerate TRUE or FALSE)
user can now select a lag for MCMC chain
new plot_matrix
function
new slfm
function which fits a model to a matrix of microarray data
new slfm_list
function which fits models to a set of matrices contained in a folder
new process_matrix
function which pre-process a set of matrices contained in a folder