Figure 9

Generalized cluster analysis applied to a protein sequence. The sequence of the rat homer evh1 domain (pdb 1ddv, chain A) is shown. Observed secondary structures, as deduced from the experimental three-dimensional structure, are shown above the HCA plot. The P-codes of the hydrophobic and loop clusters are indicated below the HCA plots and are colored when the corresponding clusters have noticeable secondary structure preferences, as reported in Additional files 3 and 5 (green: β-strands, red: α-helices, blue: coil). The corresponding frequencies of association with the preferred secondary structure (OPS and APC rules for the hydrophobic and loop clusters, respectively) are indicated below the P-codes. The small clusters, which have no clear preferences for regular secondary structures, are indicated in grey. Note that the hydrophobic cluster with P-code 49, which is associated at 41 and 36 % with α-helices and β-strands, respectively, can be suggested in a β-strand conformation, owing to its amino acid composition (three isoleucine, one threonine).