mvabund or other SIMPER alternatives

[Rosemary Hartman]

Hello all,

Happy eclipse day!

I am playing with a very messy community composition data set, and need some statistical advise. I know there are differences between my sites and habitat types (using vegan's adonis function, and NMDS plots), but I would like to know which taxa are most important in contributing to the observed difference. I have used SIMPER (similarity percentages) analysis for this in the past, but SIMPER has been criticized for conflating within-group and between-group variance. The interwebs and various journal articles suggest using multiple univariate and or/multivariate GLMs using package "mvabund". However, I can't seem to get it to work. I'm running into one error message after another that I can't interpret. Has anyone used mvabund or a similar analysis? 

I've attached some sample code and data if anyone has advice and/or I'd be happy to talk multivariate stats in person if anyone wants to chat over a beer.

Thanks,

--

Rosemary Hartman
roseh...@gmail.com
301-538-0136

[Ryan Peek]

Hi Rosemary, 

I'm not certain it's exactly what you want, but I've used the indicspecies package (https://cran.r-project.org/web/packages/indicspecies/vignettes/indicspeciesTutorial.pdf) to do some similar analysis to assess which taxa/spp are most responsible for differences between groups/sites. Another related function/package is vegan::mrpp which is helpful for testing whether the differences in community composition between two or more groups are significant using group dissimilarities. The strassoc() is useful for bootstrapping across your groups to calculate confidence intervals.

Happy to chat more if you have questions on implementation/interpretation. I found the linked tutorial (above) was pretty helpful.

Cheers,

Ryan

[Rosemary Hartman]

Thanks so much! indicspecies is great. It doesn't do exactly the same thing as SIMPER, but I think it is more informative to show which species are associated with certain groups than the contribution to differences shown by SIMPER.

Now I just need to figure out how to display the results in a way that isn't just a monstrous table of statistics.

--
Check out our R resources at http://d-rug.github.io/
---
You received this message because you are subscribed to the Google Groups "Davis R Users' Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to davis-rug+unsubscribe@googlegroups.com.
Visit this group at https://groups.google.com/group/davis-rug.
For more options, visit https://groups.google.com/d/optout.