output file format for unifrac_g beta diversity analysis

[Andy]

Hello,

I'd like to use the unifrac_g beta diversity calculation. From my understanding, this metric makes a phylogenetic tree of all OTUs present in sample i and sample j, and then calculates the fraction of branch length in this tree that is exclusive to sample i.

In the asymmetric distance matrix that is produced, is sample i across the top (columns) and sample j down the side (rows) of the table or the other way around?

Thanks,

Andy

[Greg Caporaso]

Hi Andy, 

Your interpretation is correct. I think an example will help to explain this.

If this is your OTU table:

#OTU ID s1 s2

0 1.0 1.0

1 1.0 0.0

2 1.0 0.0

3 1.0 0.0

4 1.0 0.0

and you get the following output matrix:

s1 s2

s1 0.0 0.65

s2 0.0 0.0

This should be interpreted as sample s1 having unique branch length of 0.65 relative to sample s2. s2 on the other hand has unique branch length of 0.0 relative to sample s1. 

Does that answer your question? 

Greg

[Andy]

Yes - thanks Greg!

Andy