Hierarchical Clustering In R

Question

[Sanju]

Hi all,

I have to perform hierarchical clustering in R. I tried the following code to import data from excel file and I created a diagonal matrix.

n_seq <- 15
mat <- matrix(NA, ncol=n_seq, nrow=n_seq)    
for (idx in 1:n_seq) {
  mat[idx,idx] <- 0.0
}  
for(idx in 1:(n_seq-1) ) {
  intemp <- read.xls("C://clustal.xls", sheet = idx );         
  mat[(1+idx):n_seq,idx] <- intemp[1:(n_seq-idx), 11]
}

After this , when I use hclust function I am getting the following error.

hclust(mat)
Error in if (n < 2) stop("must have n >= 2 objects to cluster") : 
argument is of length zero

How to remove this error? Please help me.

[Neilfws]

As it stands this is an R usage question, more suited to StackOverflow than here. Please indicate the relevance to bioinformatics.

[Yogesh Pandit]

Your matrix object might be null. Also make sure you have Perl installed. read.xls() runs perl in the backend.

[Michael]

This is off topic imho. It's a pure R question.

[Neilfws]

Also, check the contents of your matrix, mat. It may not contain what you think it should contain.

Answer 1

[David Quigley]

Use dist() to calculate the distance matrix. Then feed that into hclust(). Also, you seem to be working a lot harder than you have to.

1) Save your excel sheet as a tab-delimited text file called foo.txt with the format:

PROBE    P1    P2    P3
sample1    1    2    3
sample2    3    4    6
sample3    6    2    6
sample4    3    4    3

2) Call

x = read.table("foo.txt", sep='\t', header=T, row.names=1)
plot( hclust( dist( x ) ) )

[H2P]

@David I have multiple sheets in my excel file. I think, *.txt doesn't support excel file with multiple sheets. Any other methods?

Answer 2

[Mo]

If you have multiple sheets you can do as follows

library(XLConnect)
yourM <- loadWorkbook(system.file("your data.xlsx", package = "XLConnect"))
Yourdata = readWorksheet(yourM, sheet = getSheets(yourM))

First of all, try to check whether you imported your data correctly. Is it representative? If so, then you can follow below

The following is the basic to do the analysis , however, you can always play with different linkage method etc.

d <- dist(as.matrix(your data))
hc <- hclust(d)
plot(hc)

It is also a good idea to paste a part of your data so that we can import it and check whether something is wrong with your data or the way you import it is not good etc.