bioDist: bioDist
In lunacab/STATegRa: Classes and methods for multi-omics data integration

Description Usage Arguments Value Author(s) Examples

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.

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, ...)

`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

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

David Gomez-Cabrero

data(STATegRa_S1)
data(STATegRa_S2)
require(Biobase)

# 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 Gene-gene distance computed through miRNA data
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")

# Create Gene-gene distance through mRNA data
bioDistmRNA<-new("bioDistclass",
                 name = "mRNAbymRNA",
                 distance = cor(t(exprs(mRNA.ds)),method="spearman"),
                 map.name = "id",
                 map.metadata = list(),
                 params = list())

###### Generation of the list of Surrogated distances.

bioDistList<-list(bioDistmRNA,bioDistmiRNA)
sample.weights<-matrix(0,4,2)
sample.weights[,1]<-c(0,0.33,0.67,1)
sample.weights[,2]<-c(1,0.67,0.33,0)

###### Generation of the list of bioDistWclass objects.

bioDistWList<-bioDistW(referenceFeatures = rownames(Block1),
                       bioDistList = bioDistList,
                       weights=sample.weights)

###### Plot of distances.
bioDistWPlot(referenceFeatures = rownames(Block1) ,
             listDistW = bioDistWList,
             method.cor="spearman")
             
###### Computing the matrix of features/distances associated.

fm<-bioDistFeature(Feature = rownames(Block1)[1] ,
                   listDistW = bioDistWList,
                   threshold.cor=0.7)
bioDistFeaturePlot(data=fm)