fitBmeaSingle: Fit & summarise the BMEA model for a single gene(unit)
In steveped/BMEA: Bayesian Modelling for Exon Arrays

Description Usage Arguments Details Value

This function fits the BMEA model & provides the summarised output for a single gene.

1 2	fitBmeaSingle(celSet, bgCelSet, unit, conditions, contMatrix, ..., keepSims = FALSE, zGene = 4.265, zExon = 1.645)

`celSet`	an `AffymetrixCelSet` with the data to be fit
`bgCelSet`	a list with components `$lambda` & `$delta`. Each of these must be an `AffymetrixCelSet` containing the means & standard deviations for the background signal priors
`unit`	the unit (i.e. gene) to be fit
`conditions`	a vector of factors specifying which cell-type/condition each array in the `celSet` belongs to
`contMatrix`	a contrast matrix for the summarised output
`...`	used for passing variables such as `mcmcParam` to `runMCMC.BMEA`
`keepSims`	logical variable. If `TRUE` all sims from the process & contrasts will be kept
`zGene`	the zScore below which the gene is determined to be not detectable above background
`zExon`	the zScore below which the exon is determined to be not detectable above background

This function runs all the necessary checks & fits the BMEA model for a single unit (i.e. gene). The bgCelSet must be written to disk beforehand, during the preparatory stages.

The value of zGene can be set to NULL if the filtering out of low-expressed genes is not desired. This value can be changed to any value to effectively restrict the range of expression values in the dataset to those with a high signal-to-noise ratio. The value of zExon can also be set to NULL to bypass the filtering of exons which are not confidently detectable above background.

Returns a list with the following components:

$summary a matrix with the summary for all the model parameters, with convergence statistics rHat & nEff
$logFC a matrix with a separate row for each contrast. Includes mean, sd & key quantiles, with the values maxP & B. maxP refers to the maximum of the proportion of samples which were >=0 or <=0. B is calculated using the weak inequalities log(p>=0) - log(p<=0) as non-zero values for phiLogFC are possible under a (yet to be implemented) mixture model for phi.
$phiLogFC a list with a component for each contrast. Each component is a matrix as above for logFC, but with each row representing an exon.
$sims if keepSims=TRUE, this will return a list with components $data for the simulation output, $logFC for the sampled values of log fold-change and $phiLogFC for the exon-level sampled values for the change in phi. Otherwise, returns NULL.