STATegRa (version 1.0.0)

bioDist: bioDist

Description

Function to compute a bioDistclass object from profile data and a mapping. For details of the process see the user's guide, but briefly the process involves using the mapping to identify reference features appropriate to each surrogate feature (if any), aggregating the surrogate data into pseudo-data for each reference feature, and then calculating the correlation distance between the reference features according to the surrogate data.

Usage

bioDist(referenceFeatures=NULL,  reference=NULL, mapping=NULL,  referenceData = NULL,  surrogateData = NULL,                       filtering = NULL, noMappingDist = NA,  distance = "spearman",  aggregation = "sum", maxitems = NULL, selectionRule = "maxFC", expfac = NULL, name = NULL, ...)

Arguments

referenceFeatures

subset of features to be considered for the computation of the distances. If NULL then the features are first gathered from the features in referenceData. If referenceData is not provided then the list of features are gathered from mapping (bioMap class) and using the reference.

reference

A character indicating the variable that is being used as features to compute distance between

mapping

The mapping between feature types

referenceData

ExpressionSet object with the data from the reference features.

surrogateData

ExpressionSet object with the data from the surrogate features.

filtering

A filtering for the bioMap class. To be implemented.

noMappingDist

Distance value to be used when a reference feature do not map to any surrogate feature. If "max", maximum indirect distance among the rest of reference features is taken. If NA, distance weights are re-scaled so this surrogate association is not considered. If a number then the missing values are replaces with that value.

distance

Distance between features to be computed. Possible values are "pearson", "kendall", "spearman", "euclidean", "maximum", "manhattan", "canberra", "binary" and "minkowski". Default is "spearman".

aggregation

Action to perform when a reference feature maps to more than one surrogate feature. Options are "max", "sum", "mean" or "median" and the the values are aggregated according to the chosen statistic.

maxitems

The maximum number of surrogate features per reference feature to be used, selected according to "selectionRule" parameter. Default is 2.

selectionRule

Rule to select the surrogate features to be used (the number is determined by "maxitems"). It can be one of the following: (1) "maxcor" those presenting maximum correlation with corresponding main feature; in this case "referenceData" must be provided and the columns must overlap in at least 3 samples; (2) "maxmean": average across samples is computed and those features with higher mean are selected; case (3) is simmilar to (2) but considering other statistics: "maxmedian", "maxdiff", "maxFC", "sd" , "ee".

expfac

Not in use yet.

name

character that describes the nature of the bioDist class computed

...

extra arguments passed to dist, eg "p=value" for the power used if calculating minkowski distance

Value

An object of class bioDistclass containing distances between the features in surrogateData.

See Also

bioDistclass,bioDistclass-constructor

Examples

# Definition of a bioDistclass Object
data(STATegRa_S1)
data(STATegRa_S2)

# Truncate data for brevity
Block1 <- Block1[1:100,]
Block2 <- Block2[1:100,]

## Create ExpressionSets
mRNA.ds <- createOmicsExpressionSet(Data=Block1,pData=ed,pDataDescr=c("classname"))
miRNA.ds <- createOmicsExpressionSet(Data=Block2,pData=ed,pDataDescr=c("classname"))
  
## Create the bioMap  
map.gene.miRNA<-bioMap(name = "Symbol-miRNA",
                metadata =  list(type_v1="Gene",type_v2="miRNA",
                                 source_database="targetscan.Hs.eg.db",
                                 data_extraction="July2014"),
                map=mapdata)  
  
  
## Create bioDistclass Object
bioDistmiRNA<-bioDist(referenceFeatures = rownames(Block1),     
             reference = "Var1",
             mapping = map.gene.miRNA,
             surrogateData = miRNA.ds,  ### miRNA data
             referenceData = mRNA.ds,  ### mRNA data
             maxitems=2,
             selectionRule="sd",
             expfac=NULL,
             aggregation = "sum",
             distance = "spearman",
             noMappingDist = 0,
             filtering = NULL,
             name = "mRNAbymiRNA")