aclust: aclust
In peplib: Peptide Library Analysis Methods

Description Usage Arguments Details Value Note Author(s) See Also Examples

Clusters a matrix of peptide sequences using kmeans or agglomerative clustering

1 2	aclust(sDistMatrix, clusterNumber, verbose = T, type = "kmeans", knstart = 20)

`sDistMatrix`	`sDistMatrix` is a distance matrix resulting from a call to `dist(seqs)`, where `seqs` is a `Sequences` object. You may also just pass the sequences and a call to `dist` will be done automatically.
`clusterNumber`	The number of clusters to create.
`verbose`	Controls how much logging information is printed.
`type`	There are two `type` options: `"kmeans"` and `"agglomerative"`. `"kmeans"` causes `aclust` to call the normal `kmeans` method available in R. `"agglomerative"` uses a hierarchical clustering technique on the distance matrix.
`knstart`	The number of times to execute the `"kmeans"` algorithm to ensure a global minimum is reached.

During each round, the agglomerative clustering technique combines the points or sets of points which have the smallest distance to one another. The distance measure is the smallest pairwise distance between two sets of points.

aclust returns a list where each element represents a cluster. Each element in the list is an array of integers, where each integer is the index of sequence (the row of the input sequence matrix).

The "kmeans" is much much faster.

Andrew White

Sequences, kmeans

#load some example data
data(TULASequences)

##Cluster
clusters <- aclust(dist(TULASequences), 3, knstart=1)
print(length(clusters[1]))
print(length(clusters[2]))
print(length(clusters[3]))