| 1 | 2 | 3 | 4 | 5 | |
| 1 | 0 | ||||
| 2 | 1 | 0 | |||
| 3 | 5 | 6 | 0 | ||
| 4 | 5 | 6 | 1 | 0 | |
| 5 | 6 | 6 | 1 | 2 | 0 |
| 1,2 | 3 | 4 | 5 | |
| 1,2 | 0 | |||
| 3 | 5.5 | 0 | ||
| 4 | 5.5 | 1 | 0 | |
| 5 | 6 | 1 | 2 | 0 |
2. Select 3,4 as one of the pairs with minimum value. Result:
| 1,2 | 3,4 | 5 | |
| 1,2 | 0 | ||
| 3,4 | 5.5 | 0 | |
| 5 | 6 | 1.5 | 0 |
3. Now the minimum is (3,4),5. Result:
| 1,2 | (3,4),5 | |
| 1,2 | 0 | |
| (3,4),5 | 5.67 | 0 |
The distance between (1,2) and ((3,4),5) was calculated as the weighted average of the distance between (1,2) and 5 (which is 6) and the distance between (1,2) and (3,4) which is 5.5. Weighting each by the number of nodes below we get (2*5.5+1*6)/3=5.67. See equation 9.2 on page 55 in the booklet.
Note that if we had used WPGMA, the distance beween (1,2) and ((3,4),5) would simply have been (5.5+6)/2=5.75 which would effectively have given the same weight to the distance between (1,2) and 5 and that between (1,2) and (3,4) meaning that sequence 5 has more influence than each of the sequences 3 and 4.
Now the lowest pair is (1,2), ((3,4),5). The resulting tree is (edge lengths not proportional to number of mutations):