MAPs antibody structure prediction workflow. The MAPs database can be used to predict an antibody structure in three steps. First, a prototype sequence is predicted. In this work, that is accomplished by assigning germline genes that minimize the number of AA changes between the prototype and target antibody. However, prototype sequences from alternative methods that a user has confidence in can be used. In the second step, a prototype model is created by assembling and mutating the most similar models in the MAPs database. Finally, a predicted model of the antibody is created by mutating the prototype structure. The variable domain depicted for steps two and three is the VH of PDB 3ncJ  with four of the mutations highlighted.