A Method for Estimating Evolutionary Trees from Genetic Data Many methods use data on genetic differences to estimate phylogenies. It is beyond the scope of this chapter to review all of them. Instead, we shall present just one method, called the unweighted pair group method using arithmetic averages, or UPGMA. This method is not the most powerful one for estimating phylogenies from genetic data, but, in spite of its lengthy name, it is intuitively straightforward. Furthermore, UPGMA works reasonably well under many circumstances.

[Page 654] The starting point for UPGMA is a table of genetic distances among a group of species [Figure 26-19(a)]. To demonstrate the technique, we will use data from DNA hybridization studies by Charles Sibley and Jon Alquist, who computed genetic distances between humans and four species of apes: the common chimpanzee, the gorilla, the siamang, and the common gibbon.

Figure 26-19. Phylogeny reconstruction by UPGMA. (This item is displayed on page 655 in the print version) [View full size image]

The method uses the following steps:

1. Search for the smallest genetic distance between any pair of species. In Figure 26-19(a), this distance is 1.628, the distance between human and chimpanzee. Once identified, the corresponding species pair is placed on neighboring branches of an evolutionary tree [Figure 26-19(b)]. The length of each branch is half the genetic distance between the species (1.628/2), so the branches connecting human and chimpanzee to their common ancestor are each about 0.81.

2. Recalculate the genetic distances between this species pair and all other species [Figure 26-19(c)]. The genetic distance between the human-chimpanzee (Hu-Ch) cluster[*] and the other species is the average of the distances between each member of the cluster and the other species. For example, the genetic distance between the human-chimp cluster and the gorilla is the average of the human-gorilla distance and the chimp-gorilla distance, or

[*] We say "cluster" even though what we really have is a pair; the reason is that we can (and will) form larger groups and it is convenient to call them by the same, more inclusive name - ergo cluster.

3. Repeat steps 1 and 2 until all the species have been added to the tree.

The smallest distance in the recalculated table is 1.95, the distance between siamang and gibbon [Figure 26-19(c)]. To build the next section of the tree, siamang and gibbon are placed on neighboring branches, with lengths equal to half of 1.95, or 0.98. [Figure 26-19(d)].

Recalculating the table, we now find that there are two clusters plus the gorilla [Figure 26-19(e)]. The genetic distance between the human-chimp cluster and the siamang-gibbon cluster is the average of four distances: human-siamang, human-gibbon, chimp-siamang, and chimp-gibbon.

Now the smallest genetic distance is 2.239, the distance between the human-chimp cluster and the gorilla. We therefore add a gorilla branch to the tree and connect it to the common ancestor of the human-chimp cluster [Figure 26-19(f)]. The branches are drawn so that the distance between the tips of any two branches in the human-chimp-gorilla cluster is 2.239.

Recalculating the table for the final time [Figure 26-19(g)], we find that the genetic distance between the human-chimp-gorilla cluster and the siamang-gibbon cluster is the average of six genetic distances: human-siamang, human-gibbon, chimp-siamang, chimp-gibbon, gorilla- siamang, and gorilla-gibbon. This distance is 4.778, which allows us to complete our evolutionary tree [Figure 26-19(h)]. The tree indicates that humans and chimpanzees are one another's closest relatives. That is not to say that humans evolved from chimpanzees; rather, humans and chimpanzees share a more recent common ancestor than either shares with other species on the tree.