Thanks for this thread. I really enjoyed seeing the science discussed. I've been interested in computational vaccine design and HIV evolutionary dynamics for a long time. I kind of see the whole thing as a dynamical system and a hidden Markov model. Essentially since the copy mechanism is leaky and produces multiple variants, and the immune system produces a selective response, how do you know that the single epitope you vaccinate against will be sufficient? I would expect their to be multiple stable populations and hence a need multiple vaccine epitopes (including populations that are stable under specific drug regimes).

This is what makes rgejman's phage display comments so interesting. We already know two major populations (Hiv-1 and Hiv-2 exist).