compareMSI_hc: Fit hiearchical spatial model to MSI data, with improved...
In ajharry/msiCompare:

Description Usage Arguments Value

compareMSI_hc is used to fit a hiearchical Bayesian spatial model to MSI data using a Gibbs Sampler MCMC approach. The model is fit separately for each m/z feature. This version is distinct from compareMSI because it uses hiearchical centering to improve mixing of the MCMC. This function will be called by msiCompare if the MSImageSet is a multi tissue experiment with all tissues coming from the same donor or all from different donors.

compareMSI_hc(msset, conditionOfInterest, feature, nsim = 5000,
  burnin = 2500, trace = T, piPrior = 0.1, seed = 1, logbase2 = F,
  coord = NULL, type.neighbor = "radius", radius.neighbor = 1,
  maxdist.neighbor = NULL, spInit = NULL, bioRep = NULL, techRep,
  beta0 = 0, prec0 = 0.01, precAlpha0 = 0.01, d0 = 0.001, g0 = 0.001,
  rd = 1e-05)

`msset`	an object of class "MSImageSet"
`conditionOfInterest`	a vector or factor giving the level of the condition of interest for each pixel in msset
`feature`	the index of the m/z features for which the model should be fit
`nsim`	number of desired MCMC samples
`burnin`	number of MCMC samples to discard
`trace`	logical, should the full list of MCMC samples be returned for each variable?
`piPrior`	prior probability of differential abundance
`seed`	random seed
`logbase2`	logical, should the intensities be log transformed?
`coord`	data fram of coordinates of the MSImageSet, with columns 'x' and 'y'
`type.neighbor`	neighborhood type (see adj.grid)
`radius.neighbor`	desired neighborhood radius if neighborhood type 'radius' is selected (see adj.grid)
`maxdist.neighbor`	maximum distance for locations to be considered neighbors if neighborhood type 'max.dist' is selected (see adj.grid)
`spInit`	optional, provide precomputed spatial information from output of intializeSpatial
`bioRep`	optional, vector or factor giving the individual/donor to which pixel in the msset belongs
`techRep`	vector or factor giving the tissue to which each pixel in the msset belongs
`beta0`	prior mean of baseline effect
`prec0`	prior variance of baseline effect
`precAlpha0`	prior mean of condition 2 effect
`d0`	shape parameter of hyperprior of variances
`g0`	scale parameter of hyperprior of variances
`rd`	ratio of spike variance to slab variance for condition 2 effect