Hello everyone!

1. I understand that I can use unifrac.weighted to build two types of dendograms based on samples and sequences, correct?
2. Unifrac weighted includes incorporates the abundance of the OTUs and also branch lengths, right?
3. When I use unifrac.weighted command to see differences by sample I input a tree built using a dissimilarity matrix with Bray-Curtis calculator (in my case), like it is shown in the 454 SOP, right?
4. So what happens when I run unifrac.weighted, is that unifrac converts the bray curtis matrix to a unifrac distance matrix and then builds a new tree based on those distances? Is this correct?
5. Then to test for statistical significance, unifrac (with the ‘random’ option) runs monte carlo permutations (1,000) and output p scores? is this right?
6. When I open the dendograms I can see two values that appear on them some at each branch node and some on top of the branches. I am assuming that the values at the branch nodes are bootstrapping values and the ones on top of the branches correspond to branch lengths, is that correct?

thank you so much!!!
Astrid

There are no dumb questions, :). I would be happy to help.

1. I understand that I can use unifrac.weighted to build two types of dendograms based on samples and sequences, correct?

Yes, you can use mothur to do both. Pat gives examples of using the unifrac commands in the http://www.mothur.org/wiki/454_SOP Alpha and Beta diversity sections.

1. Unifrac weighted includes incorporates the abundance of the OTUs and also branch lengths, right?

http://www.mothur.org/wiki/Weighted_UniFrac_algorithm
http://www.mothur.org/wiki/Unweighted_UniFrac_algorithm

1. When I use unifrac.weighted command to see differences by sample I input a tree built using a dissimilarity matrix with Bray-Curtis calculator (in my case), like it is shown in the 454 SOP, right?

Yes, you want to run the tree.shared command and then the unifrac commands.

1. So what happens when I run unifrac.weighted, is that unifrac converts the bray curtis matrix to a unifrac distance matrix and then builds a new tree based on those distances? Is this correct?

No, the tree.shared command uses the bray curtis (or whichever calculator you choose) matrix to generate a tree that relates your samples to each other. The unifrac commands use this tree to find the diversity measurement. The unifrac commands have an option to output results in a distance matrix form.

1. Then to test for statistical significance, unifrac (with the ‘random’ option) runs monte carlo permutations (1,000) and output p scores? is this right?

Yes, you can adjust the number of permutations as well as subsample the permutations if desired.

1. When I open the dendograms I can see two values that appear on them some at each branch node and some on top of the branches. I am assuming that the values at the branch nodes are bootstrapping values and the ones on top of the branches correspond to branch lengths, is that correct?

I am not sure. You would have to post the commands you ran and a portion of the tree.

Kindly,
Sarah