How To Create Phylogenetic Trees Using The UPGMA Method

What is the UPGMA Method of Tree Building?

The UPGMA method is the simplest of the distance methods of constructing phylogenetic trees. UPGMA stands for the Unweighted Pair Group Method with Arithmetic mean and it is a remarkably simple algorithm that can be executed by hand without the need of a computer!

First I'll go over the algorithm itself and then we can do a sample problem.


UPGMA Steps

  1. First, given a distance matrix (this is just a table that lists of species and the distances between them), pick the two taxa (a species, family, class, etc.) that are most similar. This means they will have the smallest distance between them.
  2. Next, group these two similar taxa together into one combined Operational Taxonomic Unit, aka an OTU.
  3. Create a new distance matrix by taking the distances between a given taxon and the combined taxa and averaging them to get the distance from the OTU to the given taxon.
  4. Repeat this process until all taxa are joined!!

If this all sounds complicated, don't worry. Now we'll go through an example (with pictures)!


 
Dog
Fish
Bird
Dog
-
-
-
Fish
5
-
-
Bird
4
6
-
Monkey
3
7
5
Initial Distance Matrix

As you can see, the closest relationship is between the dog and the monkey with their distance of 3. Therefore, our first OTU will be formed with these two taxa.

Next, we will form a new distance matrix using this new Dog-Monkey OTU by taking the distance between the taxa and the dog and the monkey then dividing by two.

For example, the distance between the bird and the Dog-Monkey OTU would be found by taking the distance between the bird and the dog (4), the distance between the bird and the monkey (5), adding them up (9), and diving this by two to get 4.5 . Below is the resulting distance matrix.

 
Dog-Monkey
Fish
Dog-Monkey
-
-
Fish
4
-
Bird
4.5
6

As you can see, in this new distance matrix the closest taxa are the Dog-Monkey OTU and the Fish with a distance of 4. We will combine these to form a new OTU which we will call (Dog-Monkey)-Fish. The distance for this one is 4 divided by 2 which is 2!

This means it is .5 further than the Dog-Monkey OTU as you can see in the image to the right.

We will now create yet another distance matrix the same way we did previously.

 
(Dog-Monkey)-Fish
(Dog-Monkey)-Fish
-
Bird
5.25

Wasn't that easy?! Now we'll just add bird to the tree as shown below:

That's It!

See, building trees with UPGMA isn't so hard after all!



Other Methods of Making Phylogenetic Trees

Other than Distance Methods, there are two other commonly used methods of building phylogenetic trees. These are the Maximum Parsimony method and the Maximum Likelihood/Baynesian methods. I'll briefly describe both of these before we finish up.

Maximum Parsimony Method

The Maximum Parsimony method for constructing phylogenetic trees works by making a few assumptions. First, we follow the principle of Occam's Razor which simply states that simpler hypotheses are preferred to more complex ones. This is also known as the Principle of Parsimony and is just saying that the phylogenetic tree with the fewest evolutionary steps is preferred.

Second, the Maximum Parsimony method only takes into account synapomorphies, or traits that are shared by two taxa and their most recent common ancestor, but not this ancestor's ancestor!


Maximum Likelihood Method

The Maximum Likelihood and Baynesian Methods (these are not the same thing, but they function in similar ways) are complex algorithms that incorporate information about how DNA sequences evolve and mutate.


More by this Author


Click to Rate This Article
working