Hi there,
I have a dataset with 26 observations and 7039 variables each. The data looks like this
| SampleID | Time | Prot1 | Prot2 | Prot3 | Prot4 | Prot5 | Prot6 |
|---|---|---|---|---|---|---|---|
| t0_2 | Start | 0.004898333 | 0.000986609 | 0 | 0 | 0.020992855 | 0 |
| t0_3 | Start | 0.006429024 | 0 | 0 | 0 | 0.025716094 | 0 |
| t0_5 | Start | 0.003940074 | 0.000925866 | 0.000990521 | 0 | 0.015760296 | 0.001959153 |
with Prot1 - Prot7039. Missing data has been replaced with 0. To visualize and test my data, I want to conduct a NMDS using the vegan package. However, when I run the NMDS I get an error saying
*** Solution reached
Warning message:
In metaMDS(data_1, distance = "bray", k = 2) :
stress is (nearly) zero: you may have insufficient data
The stress values is usually around 9.873474e-05. These are the commands I ran
#Load the data
Exp245 <- read.table("Exp245_NMDS_Input.txt", header = TRUE)
#Create subsets for the NMDS command
data_1 <- Exp245[,3:7039]
data_2 <- Exp245[,1:2]
#Load vegan
library(vegan)
#Run the actual NMDS analysis
NMDS <- metaMDS(data_1, distance = "bray", k = 2)
I did this analysis with a similar dataset and it worked perfectly fine. I tried to transform the data hoping that this will change something but it didn't. Do you have a solution for this? Should I maybe try another distance matrix? Or is it because there are so many 0 in the dataset which makes the matrix rank insufficient?
Thank you all in advance!
Cheers,
Marlene