mclust.lm.X: dispatch to one of the implemented cluster methods....
In brentp/clustermodelr: Models on clusters of correlated data

This method implements a sorted of methyl-eQTL with a formula specified as:

1	mclust.lm.X(formula, covs, meth, X, gee.corstr = NULL, ..., mc.cores = 4)

`covs`	covariate data.frame containing the terms in formula except "methylation" which is added automatically
`meth`	a list of matrices of correlated data or a single methylation matrix
`formula`	an R formula containing "methylation"
`X`	a matrix with columns matching those in meth. n_probes X n_samples. Each row is tested by modifying `formula` so that it becomes the independent variable in the model and tested against methylation.
`gee.corstr`	if specified, the the corstr arg to geeglm.
`mc.cores`	the number of processors to use if meths is a list of matrices to test.
`...`	arguments sent to `clust.lm`

methylation ~ disease + age

each row in X is inserted into the model and tested so the model would be:

methylation ~ X[irow,] + disease + age

and the reported coefficent and p-value are from the X[irow,] covariate. This allows one to test a number of expression probes against a (number of) cluster of correlated methylation probes. Though we could also use this to test, for example a set of methylation probes against every OTU in a microbiome study. In this way, we could find DMRs related to the microbiome.

See clust.lm

brentp/clustermodelr documentation built on May 13, 2019, 5:11 a.m.