Fig. 2.

Distribution of aligned sequence pairs as a function of the agreement between observed secondary structures. Each sequence pair (query and subject detected in pdb_select_95 with the SSEARCH algorithm with an E value = 10) has been aligned using CLUSTALW (1.8) with default parameters (see Materials and Methods). The observed secondary structures were deduced from corresponding PDB files using the DSSP program (Kabsch and Sander 1983). Agreement in secondary structure was measured using the Sov parameter (Zemla et al. 1999). False positive pairs, open pale circles; true positive pairs, black crosses. Identity (A) and strong similarity as defined in CLUSTALW (B).