Distance based methods

Abstract

The fundamental prerequisite necessary to deal with all these topics is, of course, an efficient and biologically plausible way to assess sequence similarity: given an alphabet and initially just two sequences, how are we to measure in a biologically meaningful way the similarity between s and t? The answer depends, of course, on what mutational operations one perceives as biologically relevant can be performed to produce one sequence from another one. Two sequences s and t are considered to be similar if a few such operations can change s into t, and dissimilar otherwise. The problem which remains to be discussed in this context is, of course, how to assess the biological likelihood of the various operations which can be performed to change one sequence into another one and how to quantify numerically the (dis)similarity of any two given sequences.