After the protein sequences are embedded in Euclidean space, we
  further analyze the sequence space and turn to look for statistical
  regularities in the Euclidean space. Specifically, a statistical
  clustering model of the sequences is constructed.
  
  This chapter focuses on methods for clustering statistical data.  It
  starts with a very brief introduction of the notations used in
  cluster analysis, classical clustering algorithms, and validation
  schemes.  Then it describes our clustering algorithm.  By
  incorporating a cross-validation scheme, this algorithm is expected
  to construct a reliable model of the data with high probability.
  Cluster validity is measured by means of a few different statistical
  tests, which are described next.