Currently working on glycoprotein structure (CD2v) in frame of prediction modeling. Given glycoprotein has a lot of glycosylation sites and our question is could the glycosylation affect Ab binding and following interact with immune response.
1st step - build the antigenicity profiles (as hydrophylicity by Hoop Woods scale) as many as you can - it is very fast, for free and very informative (for example ProtScale tool on ExPaSy.org)
2nd step - try to find as many characteristic motifs as you can (for example using Prosite) - I have my own tool for scanning thousands of sequences using thousands of patterns in one time - I've not published it yet, so I can scan it for you, but I will not enable this tool for you today (I hope to publish it in month)
3rd step - analyze 3D structure of your protein to be sure that all find sites are enable for antibody
- you can make some more analysis (as protein-protein prediction or specific bingind sites prediction), but those 3 steps are the basis.