Hi Kendra

I had a similar problem a while back. As Jari says, you have to do the
subsetting while the object is an ordinary matrix, and then convert it back
to a "dist" object. To convert a distance matrix to an ordinary matrix, use:

mat <- as.matrix(disobject)

Then create the vector k, containing "TRUE" values for the entries you want
to select and "FALSE" values for those you don't. For example, in your case
(I think):

k <- bac_env$zone=="TX"

Then subset your matrix as Jari suggested and convert it back to a distance
matrix:

submat <- mat[k,k]
as.dist(submat)

HTH
Karen


a <- rownames(nutdisdf) %in% names(wa.comm1st)
nutdis.comm1st <- subset(nutdisdf, a == TRUE)   # keep only rows of spp
with abund records in comm1st.
nutdis.comm1st <- nutdis.comm1st[a==TRUE]   # keep only cols which match
rows above.

On Wed, Jul 31, 2013 at 7:13 AM, Jari Oksanen <jari.oksanen@oulu.fi> wrote:

> Kendra,
>
> Subset a matrix an convert to dist. You must subset both rows and columns
> in the same way, and therefore you must use subsetting index instead of a
> subset() function. Let k be a vector that is TRUE for selected items;
> as.dist(yoursymmetricmatrix[k, k]) is what you are looking for.
>
> Cheers, Jari Oksanen
> ---- alkuperäinen viesti ----
> Lähettäjä: Mitchell, Kendra
> Lähetetty:  31.07.2013, 02:38
> Vastaanottaja: "<r-sig-ecology@r-project.org>" <
> r-sig-ecology@r-project.org>
> Aihe: [R-sig-eco] subsetting lower triangle distance matrix based on
> variable in another object
>
> I have a number of dissimilarity matrices built in another program that
> I'd like to subset then run NMS.
>
> #read in distance matrix
> b_bc03 <-data.matrix(read.table("bac_final.an.thetayc.0.03.lt.ave.dist",
> row.names=1, header=T, fill=TRUE))  # creates a 736x736 double matrix
>
> #read in environmental variables
>  bac_env<-read.csv("bac_samples.csv", row.names=1, header=T)  # 736 obs.
> of 6 variables
>
> #The only way that I've figured out how to subset a lt matrix is to
> convert to dist object first
> bc_dist<-as.dist(b_bc03)  # dist[27048]
> tx_dist<-subset(bc_dist, bac_env$zone=="TX") # numeric[54316]
>
> Here is where it falls apart, the subset doesn't keep the labels, just the
> values so I can't convert it back to a distance matrix.  I've looked
> through ecodist but can't find anyway in that package to select only
> certain samples.
>
> Help please?
>
> thanks
> Kendra
>
>
>
> --
> Kendra Maas Mitchell, Ph.D.
> Post Doctoral Research Fellow
> University of British Columbia
> 604-822-5646
>
>         [[alternative HTML version deleted]]
>
> _______________________________________________
> R-sig-ecology mailing list
> R-sig-ecology@r-project.org
> https://stat.ethz.ch/mailman/listinfo/r-sig-ecology
>
> _______________________________________________
> R-sig-ecology mailing list
> R-sig-ecology@r-project.org
> https://stat.ethz.ch/mailman/listinfo/r-sig-ecology
>

	[[alternative HTML version deleted]]