Dear R and vegan package users,
I have been experiencing problems with the metaMDS function when working on a
dataset (euk) consisting of 9 "sites" (RNA extracts of 9 biofilms samples) and
340 "species" (microbial taxa based on rRNA sequences). The problem is that I
get nMDS ordinations with overlapping points, so that it looks like 6 samples
are identical, while the remaining 3 are well separated. The dataset is rather
large, and quite complex so I do not think that this is a correct
representation of dissimilarities. Every time this happens, I get the warning
message:
"In metaMDS(t(euk_red), distance = "bray") :
Stress is (nearly) zero - you may have insufficient data"
Removing relatively rare species using:
euk_red<-euk[rowSums(euk)>100,]
results in an nMDS ordination with scattered points that looks more reasonable.
However, removing even more rare species (<1000) results in a different, but
similarly uncomplex ordination and the same warning message. Changing the
distance metric to euclidean seems even more sensitive to this problem,
yielding uncomplex ordinations with almost any rare species cutoff.
The code I am using is:
euk_MDS<-metaMDS(t(euk_red), distance="bray")
ordiplot(euk_MDS, display="sites")
I have tried the arguments engine=isoMDS, and changing maxit and trymax without
noticeable effect. Generating the distance matrices separately in vegdist
yields distance matrices that look normal to my eye. It seems like a similar
problem to a recent thread in the R-help (Fabian Boetzl, Feb 2014), but my
dataset has many species that are shared between all sites. I would be happy to
provide the dataset if this may help in figuring out what the problem is. I am
using R version 3.1.0 for MacOS, coupled to R-studio (version 0.98.510).
Thanks in advance and best regards!
Mia M. Bengtsson, University of Vienna
[[alternative HTML version deleted]]
______________________________________________
R-help@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
