I will lay down the R code below
#create a matrix
> mat<- matrix(sample.int(10, size=30, replace=T), ncol=3)
> mat
[,1] [,2] [,3]
[1,] 4 7 6
[2,] 5 6 1
[3,] 4 3 1
[4,] 2 2 9
[5,] 4 5 4
[6,] 8 9 8
[7,] 5 9 7
[8,] 8 6 7
[9,] 3 8 5
[10,] 4 3 6
# column bind the mat and the sum of the rows together
> mat1<- cbind(mat, rowSums(mat))
> mat1
[,1] [,2] [,3] [,4]
[1,] 4 7 6 17
[2,] 5 6 1 12
[3,] 4 3 1 8
[4,] 2 2 9 13
[5,] 4 5 4 13
[6,] 8 9 8 25
[7,] 5 9 7 21
[8,] 8 6 7 21
[9,] 3 8 5 16
[10,] 4 3 6 13
# order the matrix by the rowSums(mat)
> mat[order(rowSums(mat)),]
[,1] [,2] [,3]
[1,] 4 3 1
[2,] 5 6 1
[3,] 2 2 9
[4,] 4 5 4
[5,] 4 3 6
[6,] 3 8 5
[7,] 4 7 6
[8,] 5 9 7
[9,] 8 6 7
[10,] 8 9 8
# compare with the order by looking at the last column
> mat1[order(rowSums(mat)),]
[,1] [,2] [,3] [,4]
[1,] 4 3 1 8
[2,] 5 6 1 12
[3,] 2 2 9 13
[4,] 4 5 4 13
[5,] 4 3 6 13
[6,] 3 8 5 16
[7,] 4 7 6 17
[8,] 5 9 7 21
[9,] 8 6 7 21
[10,] 8 9 8 25
# order the mat matrix in a decreasing order by adding a minus sign
> mat[order(-rowSums(mat)),]
[,1] [,2] [,3]
[1,] 8 9 8
[2,] 5 9 7
[3,] 8 6 7
[4,] 4 7 6
[5,] 3 8 5
[6,] 2 2 9
[7,] 4 5 4
[8,] 4 3 6
[9,] 5 6 1
[10,] 4 3 1
# label the rows into different groups
> rownames(mat)<- c("f","a","d","c","b","d","e","a","f","f")
> mat
[,1] [,2] [,3]
f 4 7 6
a 5 6 1
d 4 3 1
c 2 2 9
b 4 5 4
d 8 9 8
e 5 9 7
a 8 6 7
f 3 8 5
f 4 3 6
> rownames(mat1)<- c("f","a","d","c","b","d","e","a","f","f")
> mat1
[,1] [,2] [,3] [,4]
f 4 7 6 17
a 5 6 1 12
d 4 3 1 8
c 2 2 9 13
b 4 5 4 13
d 8 9 8 25
e 5 9 7 21
a 8 6 7 21
f 3 8 5 16
f 4 3 6 13
> groups<- factor(rownames(mat))
> groups
[1] f a d c b d e a f f
Levels: a b c d e f
# order the matrix by group first, and then within the subgroups, order by the rowSums
> mat1[order(groups,rowSums(mat)),]
[,1] [,2] [,3] [,4]
a 5 6 1 12
a 8 6 7 21
b 4 5 4 13
c 2 2 9 13
d 4 3 1 8
d 8 9 8 25
e 5 9 7 21
f 4 3 6 13
f 3 8 5 16
f 4 7 6 17
It is extremely useful when I want to order the matrix by a pre-defined way and then plot the heatmap with heatmap2 in gplots package.