Redo how we align epitopes with no exact match

Right now, we are using BLAST for linear epitopes against their parent protein and getting the best hit. For discontinuous epitopes, we are doing a residue-by-residue count. Bjoern suggested aligning the source protein to the parent and then getting sequence identity that way as it should cover more hits for linear epitopes and be more accurate for discontinous. 