We have been engaged in the research of the largest protein family since our group has been established. Numerous structure prediction methods have been developed based on protein sequence and structure analysis; such methods can be found on the CYSREDOX web server for the prediction of cystein residues participating in disulphide bonds. The concept of stabilization centers has been introduced for non-covalent crosslink playing a dominant role in protein structure stability. A new method has been developed for the prediction and identification of amino acids in these centers based on the amino acid sequence and on the protein structure, respectively (i.e. SCide, SCpred, SRide web servers). Among others, these results have been successfully applied for the analysis of the functionally regulated stability of MHC proteins and of the properties of TIM barrel structural elements, respectively, as well as for the demonstration of the divergent evolution of PD-(D/E)XK restriction endonucleases. The structural properties of several proteins have been determined by molecular mechanical and molecular dynamical tools. The role of the hydration sphere in protein – DNA recognition processes has been shown. Also, the role of metal ions in the functionality of endonucleases has been identified. Our most recent results are in the development of novel methodology for the modeling of biochemical reactions.