Alternative UPGMA & NJ phenetic analyses of Panda relationships

    The Unweighted Pair-Group Method  with Arithmetic Averaging (UPGMA) algorithm (left) assumes equal rates of evolution, so that branch tips come out equal. Note that the feliform Lynx is expected to be dissimilar to the other caniform taxa, and the UPGMA shows this.

    The Neighbor-Joining (NJ) (right) algorithm allows for unequal rates of evolution, so that branch lengths are proportional to amount of change. If rates on different branches are not markedly unequal, the branching orders produced by the two methods will not differ, as in this example. With NJ, placement of Lynx is not automatic, and it must be specifically placed as the outgroup to the caniform taxa.

Text material © 2024 by Steven M. Carr