0
$\begingroup$

I would like to make a NMDS with biomass of different prey groups in stomach content of fish.

I have already made one where the data matrix consists of 0 and 1, and this one went fine but are not able to do it when adding the actual weights.

I have used the MASS and vegan package. When using binomial data matrix:

mymetaMDS <- metaMDS(comm, distance = "jaccard", na.rm = TRUE, binomial = TRUE) 

and this works fine.

But when I try to create the dissimilarity matrix with the weight data matrix:

mymetaMDS <- metaMDS(comm, distance = "jaccard", na.rm = TRUE, binomial = FALSE) 

Error in if (any(dist < -sqrt(.Machine$double.eps))) warning("some dissimilarities are negative -- is this intentional?") : missing value where TRUE/FALSE needed In addition: Warning messages: 1: In distfun(comm, method = distance, ...) : you have empty rows: their dissimilarities may be meaningless in method "jaccard" 2: In distfun(comm, method = distance, ...) : missing values in results

Any suggestions any one?

$\endgroup$
  • $\begingroup$ Why apply non-metric methods to measured data? More generally, it is best to assume ignorance of biology when posting here. People don't necessarily know how biomass is measured. $\endgroup$ – Nick Cox Feb 6 '14 at 17:29
1
$\begingroup$

I would recommend trying to use Bray-Curtis (distance= "bray") instead of Jaccard distance as your similarity matrix. It looks like the jaccard distance is really only useful for binary data (presence/absence) while The bray-curtis matrix has been found to be robust for many abundance type data sets, especially those with many paired zeros like stomach content data can have.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.