Analysis of microbial community data collected from iPad swabs

Question

We have taken a number of swabs (~200) from iPads used in our clinic, and after the first half we've introduced an intervention to disinfect all iPads after every use (in addition to the standard hygiene measures). Besides doing culture, we've sequenced DNA to obtain information about what different bacterial species were contained in every sample. The sequencing data now consists of ~120 species, given as percentage contained in each sample (eg. sample5 contains species1 40%, species7 30% and species18 10%).

As an additional analysis I would now like to compare the microbiome (i.e. abundance of species from each sample) before the intervention with that after the intervention, and have thought of creating a (dis)similarity matrix of all species and using multidimensional scaling before and after intervention.

Is this an adequate way to do it? Are there other/better methods?

As I've never done this, and the resources on the web seem a bit scarce, do you have any pointers to recommended literature or tutorials, especially for Stata 13?

Answer 1

That sounds like what people do these days. There are tons of ecological similarity measure that you can use in order to construct your similariy metrics. you can use spectral clustering which are special case of multidimensional scaling for similarit matrices. You can use pair-wise test to see level of significants; I would say the best source you can start with is the phyloseq package;