[R-sig-eco] The problems on BIO-ENV procedure by [R]

[Dave Roberts]

Shun Tsuobi,

    isoMDS (and routines based on it) will not allow zero dissimilarity, 
which implies perfect replicates.  One alternative is to remove one of 
the replicates, but that may have unsatisfactory effect on further 
analyses.  Alternatively, if you are sure you want to keep the duplicate 
plots you can change the zero to a small value

d[d==0] <- 0.0001

and run the isoMDS on the resulting revised dissimilarity matrix.

Undoubtedly someone from the vegan group will respond to your second 
question.

Dave Roberts

On 12/05/2011 01:28 AM, 坪井　隼 wrote:
> Dear Madam / Sir,
>
> I have two questions for the use of the “R” program for the ecological
> research. I am studying the relationship between the community
> structures of environmental microbes and some environmental conditions.
> For this objective, I have known that the BIOENV procedure, which was
> developed by Clarke&  Ainsworth (1993), is available on the “R” software.
>
>
> Fist question;
> I attempted the use of the procedure to analyze the relationship between
> the variation of the microbial community structures and the
> environmental factors. However, I can not analyze the relationship based
> on isoMDS function. The isoMDS was inacceptable for my dataset. The
> command for the BIOENV procedure, which I programmed, and the error
> massage I gained was as follow;
>
>> library(MASS)
>> library(vegan)
>> communitydat<-read.table("C:/Documents and Settings/shuntsuboi/desktop
> /bray.txt", head- er=T)
>> environdat<-read.csv("C:/Documents and Settings shuntsuboi/desktop/ev.
> csv",header=T)
>> env<-environdat[,c("variablesA","variablesB.","variablesC","variablesD
> ","variablesE")]
>> d<- vegdist(communitydat, "bray")
>> isoMDS(d)
> error   isoMDS(d) : zero or negative distance between objects 1 and 2
>
> As mentioned above, I can not run the program because the error, which
> is “isoMDS(d) : zero or negative distance between objects 1 and 2”,
> occurred. What are the ways to solve this problem ? On isoMDS function,
> what are the ways that the zero distance of “Bray-Curtis distance” is
> acceptable in the function ?
>
> Second question;
> 	Based on the command as above, I ran the metaMDS function. However,
> although I could automatically describe the two dimensional ordination
> plot figure, I could not gain the X and Y value of the respective plots.
> Then, the error massage was shown as follow;
> “In ordiplot(x, choices = choices, type = type, display = display,  :
> Species scores not available”
>
> What are the ways to solve this problem ?
>
> Sincerely yours,
> Shun Tsuboi
>
> _______________________________________________
> R-sig-ecology mailing list
> R-sig-ecology at r-project.org
> https://stat.ethz.ch/mailman/listinfo/r-sig-ecology

-- 
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
David W. Roberts                                     office 406-994-4548
Professor and Head                                      FAX 406-994-3190
Department of Ecology                         email droberts at montana.edu
Montana State University
Bozeman, MT 59717-3460