Bio-Rad Quantity One 1-D Analysis Software - Manual (Page 224)

Chapter 10. Reports
10-15
The report window for a phylogenetic tree is slightly different than the standard report
window. The tree appears in a window with scroll bars for moving up and down and
left and right.
To print the tree, click on the Print button.
Click on the Cluster Method button to select a different clustering method for
displaying the tree. This will automatically recalculate and redisplay the tree.
Neighbor Joining
This type of phylogenetic tree is computed is based on minimizing the total branch
length at each stage of clustering. The method also finds branch lengths between
nodes. The approximate distance between any two samples in this tree can be found
by adding the branch lengths that connect the samples.
Other Methods
1
,
2
The following methods are based on the algorithm below:
1. Begin with n clusters—one cluster for each sample.
2. Compute the similarity matrix for the samples.
3. Convert the similarity matrix into a distance matrix d using the appropriate
distance formula.
4. Join the two clusters with the minimum distance into one cluster. Compute the
similarity value for this cluster.
5. Recompute the distance matrix d using the cluster that was formed in Step 4.
Steps 4 and 5 are repeated until there is only one cluster. The difference between the
methods below is based on the definition of minimum distance in Step 4, and on the
1. For a complete explanation of the calculations and assumptions used
to generate these dendrograms, please refer to Sneath and Sokal.
Numerical Taxonomy, San Francisco: W. H. Freeman & Company,
1973.
2. Vogt and Nagel, Clinical Chemistry 38 (2): 182-198, (1992).