[R] Compare two distance matrices

[Mattias de Hollander]

Hi all,

Thanks for the quick response. I see the ade4 package in not needed
for distance matrix computation, but as far i can see you need it for
comparing two distance matrices. In the stats package i can't find any
similiar functions like mantel.randtest or RVdist.randtest of the ade4
package. So i think this package is still needed if i would like to
make a scatter plot of the matrices. Or should i manualy compare these
matrices with a loop for example and make a plot of this?

On 10/6/05, Martin Maechler <maechler at stat.math.ethz.ch> wrote:
> >>>>> "bady" == bady  <bady at univ-lyon1.fr>
> >>>>>     on Thu, 06 Oct 2005 14:39:27 +0200 writes:
>
>     bady> Hi, hi all,
>     >> I am trying to compare two distance matrices with R. I would like to
>     >> create a XY plot of these matrices and do some linear regression on
>     >> it. But, I am a bit new to R, so i have a few questions (I searched in
>     >> the documentation with no success).
>     >> The first problem is loading a distance matrix into R. This matrix is
>     >> the output of a the Phylip program Protdist and lookes like this:
>     >> I tried with the scan() function to load the files, but with no
>     >> success. How should i load in these files? ....
>     >>
>
>     bady> you can separately load each matrix with two text files.
>
>     bady> require(ade4)
>     bady> mat1 <- read.table("mat1.txt")
>     bady> nam1 <- mat1[,1]
>     bady> mat1 <- mat1[,-1]
>     bady> row.names(mat1) <- names(mat1) <- nam1
>     bady> mat2 <- read.table("mat2.txt")
>     bady> nam2 <- mat2[,1]
>     bady> mat2 <- mat2[,-1]
>     bady> row.names(mat2) <- names(mat2) <- nam2
>
>     bady> dist1 <- mat2dist(mat1)
>     bady> dist2 <- mat2dist(mat2)
>
> but I don't see why you would need an extra package "ade4" and
> its "extra - function"  mat2dist().
>
>
> when the 'stats' package already provides the function
> as.dist(.)  {the help page of which was mentioned by the
> original poster}.
>
>
> Here is a reproducible example showing how I think as.dist()
> works sufficiently:
>
> > (m <- toeplitz(round(rnorm(6),2)))
>       [,1]  [,2]  [,3]  [,4]  [,5]  [,6]
> [1,] -0.42 -0.78 -0.42 -2.24  0.74  1.31
> [2,] -0.78 -0.42 -0.78 -0.42 -2.24  0.74
> [3,] -0.42 -0.78 -0.42 -0.78 -0.42 -2.24
> [4,] -2.24 -0.42 -0.78 -0.42 -0.78 -0.42
> [5,]  0.74 -2.24 -0.42 -0.78 -0.42 -0.78
> [6,]  1.31  0.74 -2.24 -0.42 -0.78 -0.42
> > as.dist(m)
>       1     2     3     4     5
> 2 -0.78
> 3 -0.42 -0.78
> 4 -2.24 -0.42 -0.78
> 5  0.74 -2.24 -0.42 -0.78
> 6  1.31  0.74 -2.24 -0.42 -0.78
> > ## it also works for data frames {if really needed}:
> > dm <- as.data.frame(m)
> > as.dist(dm)
>       1     2     3     4     5
> 2 -0.78
> 3 -0.42 -0.78
> 4 -2.24 -0.42 -0.78
> 5  0.74 -2.24 -0.42 -0.78
> 6  1.31  0.74 -2.24 -0.42 -0.78
> >
>
>