Hi, a general maybe fuzzy question:

How comes that distances of sequences to one consensus sequence are typically longer than when you compare them pairwise and say build a neighbour joining tree and mesure the distance from root to position of sequence in the tree

Any idea if this observation is somehow explainable?

I want to compare a (large set of sequences) and apply two different methods:

Method A calculate a consensus sequence for the set, (or form subsets by any criteria and then calculate one consensus sequence each subset) measure the distance of some sequence to the consensus-sequence (Jukes Cantor for instance)

Method B compare all the sequences against each other and build a distance matrix and then a neigbor joining tree from it calculate distance to the same sequence as before as the distance from tree root

My observation: The tree distance is shorter in most cases.

How comes?