Best Method for Comparison of Bacterial Community Compositions


I have curated 454 pyrosequencing data using Mothur. I have clustered reads from each sample into OTUs at a unique level. (Please see attached image of Excel spreadsheet)

I would like to compare the bacterial communities within these samples in a pairwise manner (taking into consideration both presence/absence as well as magnitude) in order to assess whether the bacterial communities are significantly different to one another. I am a complete novice with Statistics and have attempted only ANOVA and T-tests but as far as I understand, those are more suited to replicate results per sample rather than what I'm working with. Literature on the matter is very scant on how they actually performed the tests and I've been struggling to wrap my brain around which method is best for three days now, with no luck!

Any suggestions on how best to tackle this data would be hugely appreciated!