PCAN (version 1.0.2)

hpSetCompBestMatch: Best matches between two sets of HP terms

Description

This function returns the best matches from a semantic similarity matrix.

Usage

hpSetCompBestMatch(hpSetComp, direction = c("b", "r", "c"))

Arguments

hpSetComp

a matrix of semantic similarities between couples of HP terms

direction

taken into account. "r": best match per row. "c": best match per column. "b" (symetric): best match for the whole matrix

Value

A data frame with the compared term, the best match and the value of the match.

See Also

compareHPSets and hpSetCompSummary

Examples

## Prerequisite
data(geneByHp, hp_descendants, package="PCAN")
geneByHp <- unstack(geneByHp, entrez~hp)
ic <- computeHpIC(geneByHp, hp_descendants)

###########################################
## Use case: comparing a gene and a disease
data(traitDef, geneDef, hp_ancestors, hpDef, package="PCAN")
omim <- "612285"
traitDef[which(traitDef$id==omim),]
entrez <- "57545"
geneDef[which(geneDef$entrez==entrez),]
## Get HP terms associated to the disease
data(hpByTrait, package="PCAN")
hpOfInterest <- hpByTrait$hp[which(hpByTrait$id==omim)]

## Get HP terms associated to the gene
hpByGene <- unstack(stack(geneByHp), ind~values)
geneHps <- hpByGene[[entrez]]
## Comparison of the two sets of HP terms
compMat <- compareHPSets(
    hpSet1=geneHps, hpSet2=hpOfInterest,
    IC=ic,
    ancestors=hp_ancestors,
    method="Resnik",
    BPPARAM=SerialParam()
)
## Get the symmetric semantic similarity score
hpSetCompSummary(compMat, method="bma", direction="symSim")
bm <- hpSetCompBestMatch(compMat, "b")
hpDef[match(c(bm$compared, bm$candidate), hpDef$id),]