concordance: Evaluate the concordance between Phenotype objects

In gespeR: Gene-Specific Phenotype EstimatoR

Description

Measures include the correlation (rho) between pairs of phenotypes for the same gene, the rank biased overlap (rbo) of the top and bottom of ranked lists, and the Jaccard index (J) of selected genes.

Usage

concordance(..., min.overlap = 10, cor.method = "spearman", rbo.p = 0.98,
  rbo.k = NULL, rbo.mid = 0, uneven.lengths = TRUE)

Arguments

...	The phenotypes to be evaluated for concordance
min.overlap	The minimum number of overlapping genes required
cor.method	A character string indicating which correlation coefficient is to be computed
rbo.p	The weighting parameter for rank biased overlap (rbo) in [0, 1]. High p implies strong emphasis on top ranked elements
rbo.k	The evaluation depth for rank biased overlap extrapolation
rbo.mid	The mid point to split a ranked list, e.g. in order to split positive and negative scores choose mid=0
uneven.lengths	Indicator if lists have uneven lengths

Value

A concordance object with the following elements:

pair.test	Indicator of compared phenotypes
cor	The correlation between pairs of phenotypes for the same gene
rbo.top	The rank biased overlap of genes evaluated at the top of the ranked list
rbo.bottom	The rank biased overlap of genes evaluated at the bottom of the ranked list
jaccard	The Jaccard index of selected genes

Author(s)

Fabian Schmich

See Also

Phenotypes

plot.concordance

rbo

Examples

data(stabilityfits)
conc <- concordance(gsp(stabilityfits$A), gsp(stabilityfits$B),
gsp(stabilityfits$C), gsp(stabilityfits$D))
plot(conc)

gespeR documentation built on Nov. 8, 2020, 5:35 p.m.