A major obstacle in the development of an HIV vaccine is the
formidable mutational capacity of the virus. This rapid rate of
mutation allows the virus to quickly adapt to the adaptive immune
response, making priming of the immune system via vaccination an as
yet futile effort. In this talk, I will present some graphical models
we have developed for inferring patterns of immune escape in HIV
evolution. These patterns turn out to be remarkably consistent,
revealing details of the the immune response against the virus as well
as evolutionary constraints within the virus. Using these apparent
constraints on viral evolution, we are beginning exploration of new
vaccine candidates and hope to better understand the underlying
mechanism of the disastrous Merck vaccine failure this past September.