Dear colleagues,
I'm having troubles in some data analysis.
I have Presence/Absence data from a few communities.
I've run a nMDS based on Jaccard index matrix using vegan package in R.
Further, I'd like to know if is possible to conduct a SIMPER like analysis in order to determine which species are contributing to the similarity/dissimilarities between communities.
I've read some discussions and Jaccard index are not used in SIMPER analysis.
There is any other analysis which I could perform that would give me this answer ??
Thanks for you attention.
If you have presence/absence (binary) data you can, indeed, use a resemblance coefficient like Jaccard. You could calculate the decomposition of Jaccard between 2 samples into contributions from individual species, I suppose. What most people do in my experience is to use the Sorensen (or Dice) coefficient. This is simply the Bray-Curtis measure applied to binary data. Then SIMPER can be run as usual. The 'abundances' in the output will then be relative frequencies of occurrence. If you are using rank-based multivariate analyses then Jaccard and Bray-Curtis will give the same results as they are monotonically related to each other.
- Badges
- Science topic
- Similar topics
- Phytochemistry
- Extracts