There is technology available to identify human helper T cell epitopes. Its a computer-based method for predicting the binding of peptides to human MHC class II molecules. Computer models of MHC class II are generated with reference to known structures solved by X-ray crystallography. Individual peptides are then analysed for a preferred conformation for binding within the peptide binding groove of each MHC molecule. My question is : how these peptides are generated? based on a known antinody sequence or randomly chosen?