wallace: Computes Wallace scores comparing two clustering methods.

Description Usage Arguments Value Author(s) Examples

Description

Given two clusterings A \& B we can calculate the likelihood that two elements are in the same cluster in B given that they are in the same cluster in A, and vice versa.

Usage

1
wallace(v1, v2)

Arguments

v1

SimpleIntegerList object (output from makeClusters or makeClustersFF).

v2

SimpleIntegerList object (output from makeClusters or makeClustersFF).

Value

Vector of length 2 giving conditional likelihoods.

Author(s)

Thomas J. Hardcastle

Examples

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
# using likelihood data from a Bayesian analysis of the data

# load in analysed countData object
data(cD.ratThymus, package = "clusterSeq")

# estimate likelihoods of dissimilarity on reduced set
aM <- associatePosteriors(cD.ratThymus[1:1000,])

# make clusters from dissimilarity data
sX <- makeClusters(aM, cD.ratThymus[1:1000,], threshold = 0.5)

# or using k-means clustering on raw count data

#Load in the processed data of observed read counts at each gene for each sample. 
data(ratThymus, package = "clusterSeq")

# Library scaling factors are acquired here using the getLibsizes
# function from the baySeq package.
libsizes <- getLibsizes(data = ratThymus)

# Adjust the data to remove zeros and rescale by the library scaling
# factors. Convert to log scale.
ratThymus[ratThymus == 0] <- 1
normRT <- log2(t(t(ratThymus / libsizes)) * mean(libsizes))

# run kCluster on reduced set.
normRT <- normRT[1:1000,]
kClust <- kCluster(normRT, replicates = cD.ratThymus@replicates)

# make the clusters from these data.
mkClust <- makeClusters(kClust, normRT, threshold = 1)


# compare clusterings
wallace(sX, mkClust)

tjh48/clusterSeq documentation built on May 31, 2019, 3:40 p.m.