Please, provide:

-head of your datafram

-The command you typed to get LC50 object

I used this :

pvclust(t(gel), method.hclust="average", method.dist="euclidean")

with gel a binary data fram of presence/absence

Please note that pvclust is taking the data in transposed way according to standard convention.

Oh, sorry. this is the data and the code:

LC50 LC50$X LC50

lambda.cyalothrin deltamethrin cyfluthrin permethrin

1 0.581 0.793 1.213 2.920

2 0.750 0.137 1.340 4.535

3 0.931 0.771 1.108 11.948

4 0.516 0.351 2.635 3.703

5 0.372 0.164 0.309 4.809

6 0.288 0.197 0.413 2.469

7 0.599 0.784 2.132 8.223

8 0.059 0.022 0.045 1.218

9 0.100 0.150 0.120 2.308

> library(pvclust)

> clust.LC50

So, many things to say and ask…

-When testing try nboot=100 unless you use 10 000 in order to justify your coffee/cigaret/beer break

-When giving a part of data use head()

head(LC50)

Then now… the funniest… just try :

clust.LC50

Haha, okay, okay, is faster, but there is not another implication to use a 10 000 bootstrap? I read that is more accurate to use a 10 000 bootstrap.

At the end, I have the same error :/

"Warning message lost

In a$p[]

Use a long methods when you are sure it is working… of course 10 000 is more accuarate… if your computer is very fast and you have some time to waste just run 1 000 000 :p

so have you plot?

so… I guess 1, 2 , 3 ,… are the stuff you want to cluster, no?

just plot… I quite sure what is trying to cluster is lambda.cyalothrin deltamethrin cyfluthrin permethrin

I already have this error :

"1:inappropriate distance matrices are omitted in computation: r = 0.444444 "

In my case, it was NA values, which prevented the distance calculation.

Well at the end I used this command

library(pvclust)

CLUSTER.KC50

For my paleoecological data there were error masseges using pvclust looked like above (1:inappropriate distance matrices are omitted in computation...etc..) but the solution clust.G3