Hi Sarah,

Vegan might be quite useful for you. I have not used it myself for what you want, but I am aware of people who have and found it useful. It does contain a lot of the smaller packages which might be of more use. diversity might also be of help. If you have RStudio, that has a good package search section, where you can easily access all the help info for each package.

Hope that helps a bit? I am still relatively new to R so, I apologies if I am telling you what you already know.

Package sna or igraph can be helpful even they are more designed for social network analysis, but very flexible.

It really depends on what you want to do. igraph is much better in data handling, sna (or statnet) has more analytical procedures build in. But your projects sounds very interesting. Is there any chance to learn more about it?

Some packages that may be handy when you come across some environmental data analyses:

Package AICcmodavg (based on Akaike's information criterion)

Package mixdist (Mixture Distribution Models)

Package mvtnorm (Computes multivariate normal and t probabilities)

Package Rcmdr (friendly graphical user interface)

Multivariate Analysis of Ecological Data (Book, 2014) by Michael Greenacre and Raul Primicerio. 

Hi Sarah,

If you wannt to cross-reference data (abiotic and biotic for esample), vegan might be quite useful indeed. It offers a large panel of analysis : exploratory ones (rda, cca...), statistical using distance matrix (clustering)...

Rpart package is also very useful for ecological studies if you have a priori group.

two more options that have not been mentioned yet, although I am not sure they provide exactly what you need:

bipartitte (Dormann et al)

cheddar (Hudson et al)

cheers

Gregor

Package sna was very helpful. For great visualisations and some very basic analysis have a look at the open source Gephi. Hal Whitehead has published numerous papers and a book called 'Analysing Animal Societies' that explains things very clearly and deals with occurance  type data handling. You might be best off manipulating your two data sets into one in R before you proceed, depending on your data. Good luck!

I found igraph to be the most user friendly. Good luck!

ADE-4 is an interessant sofware

These two papers include R code in the supplemental material that will get you started running network analysis in R:

http://journal.frontiersin.org/article/10.3389/fenvs.2014.00010/abstract

http://journal.frontiersin.org/article/10.3389/fmicb.2014.00358/full

I don't believe there's a link to it in the paper, but Williams has also established a github repository with the scripts he used: https://github.com/chnops/co-occurrence

Thank you all very much for your answers! I've certainly got lots to follow up now :-) I'm planning to analyse patterns of co-occurance between fungal and bacterial taxa in my samples, hence the two datasets. I have just started using RStudio, but not come accross the package search yet! I'll be sure to look up the suggested packages.

For ecological network analysis, you can go through free software like ECOPATH, ECONET, NTWRK. In google search you will get the website and will be able to download. All are absolutely free.

Thanks

Santanu Ray

Visva-Bharati University

For the R software, "vegan" and "ade4" are good packages for community analyses, especially with presence/absence data across samples.

I think your best shot within R is the "enaR" package. This package is specifically designed to analyze ecological networks. Alternatively you might consider a structural equation modelling approach using the "lavaan" or "sem" packages.

Sarah-- I just discovered your question.  I appreciate Dr. Weterings recommendation of our enaR package.  It is designed to be a user friendly way to apply ENA to an ecosystem type network model.  Here is a link to the package vignette (https://cran.r-project.org/web/packages/enaR/vignettes/enaR-vignette.pdf) and a recent paper (http://onlinelibrary.wiley.com/doi/10.1111/2041-210X.12282/abstract).  Please let me know if you have questions or recommendations for the package.  

Another new-ish one (2020): hmsc. A hierarchical modelling of species communities. Includes phylogeny, traits, co-occurence etc. A Bayesian approach. Made need a bit of time to get acquainted with but worth it for great results!