Description Usage Arguments Details Value Author(s) Examples
Integrative Sparse KMeans
1 2 3 |
d |
combined data matrix n*J, where n is number of subjects, J=J1+J2+... and J1 is number of features in omics dataset 1 and J2 is number of features in omics dataset 2... |
K |
number of clusters |
gamma |
Penalty on total number of features. Larger gamma will yeild small number of selected features. |
alpha |
balance between group sparsity and individual sparsity. alpha=1 yeilds no group sparsity. alpha=0 yeilds no individual penalty. |
group |
Prior group information. Potentially these group can contain overlap features. group is a list and each element of the list is feature index. |
nstart |
Number of initials for Kmeans for sparse Kmeans |
wsPre |
Initial feature weight. |
penaltyInfo |
only for the purpose of gap statitics. Here we will fix the penalty design to perform gap statistics. The input should be a list of groupInfo. See groupInfo for details. |
sparseStart |
Use Sparse Kmeans to do initialization. |
silent |
Output progress. |
maxiter |
Maximum numbre of iteration between ws and Cs. |
Integrative Sparse KMeans, integrating multiple omics dataset, using prior group information.
m lists, m is length of gamma parameters. Each list is consisting of (ws=ws, Cs=Cs, objective=ADMMobject$objective, BIC=BIC, gamma=agamma,alpha=alpha
ws |
weight for each feature. Zero weight means the feature is not selected. |
Cs |
Cluster Assignment |
objective |
objective value |
obj0 |
sum of weighted separation ability. This term is for the purpose of gap statistics. |
groupInfo |
a list containing group design, alpha, gamma |
Caleb
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 | set.seed(123)
# Generate two random omics datasets
mu <- c(-3,1,3)
Simu1_mRNA <- rbind(cbind(matrix(rnorm(40*5, mu[1], 0.1),40,5),
matrix(rnorm(40*5, mu[2], 0.1),40,5),
matrix(rnorm(40*5, mu[3], 0.1),40,5)),
matrix(rnorm(10*15,0,0.1),10,15))
mu <- c(1,3,-3)
Simu1_methyl <- rbind(cbind(matrix(rnorm(40*5, mu[1], 0.1),40,5),
matrix(rnorm(40*5, mu[2], 0.1),40,5),
matrix(rnorm(40*5, mu[3], 0.1),40,5)),
matrix(rnorm(10*15,0,0.1),10,15))
## feature modules across two datasets
group <- list(c(1:10,51:60), c(11:20,61:70), c(21:30,71:80), c(31:40,81:90))
DList <- rbind(Simu1_mRNA, Simu1_methyl)
dim(DList)
## colSums(module)
K=3
nstart=20
silent=FALSE
maxiter=6
centers=NULL
error=1e-4
sparseStart = TRUE
gamma = 0.2
alpha = 0.5
d <- t(DList)
iRes <- ISKmeans(d, K=3, gamma=0.5, alpha=0.5, group=group)
|
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.