The first rational models of the complete combining site that generated 'rules' were built in 1986 (de la Paz et al, 1986; Chothia et al 1986). By 1987-1989, canonical structures for CDRs had been refined by Chothia and Lesk while Martin, Cheetham and Rees combined knowledge based and ab initio methods developed by Bruccoleri and Karplus to produce CAMAL (Combined Antibody Modeling Algorithm; Martin et al 1989).
During the 1990's many improvements were made to the canonical methods, including rules for modeling CDR H3 by the group of Nakamura (Shirai et al 1999; Kuroda et al 2008, 2009) while conformational search and other methods based on energy calculations increased the reliability of side chain placement. Whitelegg and Rees had extended the combined approach by defining many more canonical classes while retaining the conformation search methods for 'difficult' CDRs (Whitelegg & Rees, 2004). These new developments were added to the web-based program,WAM (web antibody modeling) that had been launched in 2000. In 2008 Anna Tramontano launched the PIGS (Prediction of Immunoglobulin Structure) web modeling site followed a year later by the Rosetta software from the group at Johns Hopkins.
More details of the history of antibody modeling can be found in the book "The Antibody Molecule: From Anti-toxins to Therapeutic Antibodies"©, Anthony R Rees, Oxford University Press, 2014.
Further developments in recent years, particularly by Andrew Martin with development of the abYsis databases for expert users and a rapid modeling method (abyMod) that will be launched later in 2014, have further solidified the antibody analysis and prediction process (Martin, 2013).
Our consulting group makes use of WAM and works closely with expert partners for its analysis and modeling. Structural analysis by x-ray crystallography is provided through a close relationship with contract structure determination partners providing fast and accurate structures of Fab molecules and/or their complexes with cognate antigen.