findOutliers: Identify outlier objects given a square distance matrix.

In clstutils: Tools for performing taxonomic assignment.

Description

Outliers are defined as elements with edge length to the centermost element > cutoff. The distance threshold (cutoff) can be either specified, or calculated as a quantile of all pairwise distances in the matrix.

Usage

findOutliers(mat, quant, cutoff)

Arguments

mat	square matrix of distances
quant	given all pairwise distances x, calculate distance threshold as quantile(x, quant). Values closer to 0 are more stringent.
cutoff	an absolute cutoff overriding quant

Value

Returns a boolean vector corresponding to margin of mat; outliers have a value of TRUE.

Author(s)

Noah Hoffman

Examples

library(ape)
data(seqs)
data(seqdat)
dmat <- ape::dist.dna(seqs[seqdat$tax_name == 'Enterococcus faecium',],
        pairwise.deletion=TRUE, as.matrix=TRUE, model='raw')
summary(dmat[lower.tri(dmat)])
outliers <- findOutliers(dmat, cutoff=0.015)
table(outliers)

Example output

Loading required package: clst
Loading required package: rjson
Loading required package: ape
    Min.  1st Qu.   Median     Mean  3rd Qu.     Max. 
0.000000 0.003358 0.005975 0.011041 0.013277 0.144397 
outliers
FALSE  TRUE 
  103     8 

clstutils documentation built on Nov. 8, 2020, 5:23 p.m.