MBXUCL: R Routines for metabolomic data

Description Usage Arguments Value Author(s) Examples

Unsupervised clustering on (GPL) binary presence/absence of peaks based on Ward algorithm. Calculation of MIC statistical criteria of clustering quality: Dunn, Davies-Bouldin, Rand and adjusted-Rand indexes.

1	binClustMIC(Positions, Distance, nClust, Trcl, Dendr = TRUE)

`Positions`	A binary numeric matrix
`Distance`	Choice of the distance measure adapted to binary objects ('Jaccard' or 'Ochiai')
`nClust`	The number of groups to retrieve (donors, mixtures, ...).
`Trcl`	The real groups' memberships of the samples, true class.
`Dendr`	Logical argument (TRUE/FALSE) to obtain graphical dendrogram based on the Ward algorithm.

A list of MIC quality indexes (Dunn, Davies-Bouldin, Rand and adjusted-Rand):

DunnW: Dunn index for Ward clustering
DBW: Davies-Bouldin index for Ward clustering
RandW: Rand index for Ward clustering
AdjRandW: Adjusted Rand index for Ward clustering

Baptiste Feraud

1
2
3

Pos = t(GPL[, seq(4,dim(GPL)[2], 2)])
Trcl = c(rep(1,8), rep(2,8), rep(3,8))
binClustMIC(Positions = Pos, Distance = 'Jaccard', nClust = 3, Trcl, Dendr = TRUE)