Hi all,

The problem I am facing is finding proteins that may react with antibodies that are produced against a specific protein (which I will refer to as protein X).

I was thinking of collecting all proteins that are (roughly) similar to protein X using PSI-BLAST, and the predicting linear epitopes in protein X and the BLAST hits using BepiPred (or something similar). If I will find linear epitopes with the the same amino acid composition on protein X and on one of the BLAST hits, then this would be a candidate.

Does this workflow sound reasonable?

Is there any way to incorporate structure in this problem? Protein X has a structure and I can predict discontinuous epitopes on it using programs such as DiscoTope. Even if I limit myself to search against proteins that can be found in PDB, how do I search for similar epitopes on other proteins?

Thank you very much

Gil Hornung