Vaccines to protect against rapidly mutating viruses, such as influenza and HIV, face the challenge of inducing immunity against a broad range of viral variants. Current influenza virus vaccines protect against only a narrow range of such variants, and on-going revision of the vaccine is necessary, along with annual revaccination. Some individual subjects make antibodies with considerable breadth of infectivity neutralization, but these responses do not dominate the usual repertoire. Can we understand the development of immunity — especially the molecular details of affinity maturation — well enough to design immunogens that elicit broad responses more reliably? We seek to answer that question together with collaborators at Duke University and the Duke Human Vaccine Institute, the Ragon Institute of Harvard, MIT and MGH, Boston University School of Medicine, and the University of Texas, Austin. Cell sorting and high-throughput DNA sequencing technologies allow one to derive a large number of sequences for antibodies (B-cell receptors) expressed in an individual subject at any chosen time after immunization or infection. Many of these B-cells turn out to be members of one of a small number of clonal lineages, each of which derives from a distinct naïve response. The richness of sequence information permits us to derive the amino-acid sequences of that naïve response (the “unmutated common ancestor” or UCA) and of the various intermediates at the branch points of the tree relating mature antibodies to the UCA. We can thus examine protein evolution in real time (and in a real person), expressing members of selected lineages and visualizing by crystallography and electron microscopy the molecular adjustments that accompany affinity maturation. A longer term goal is to use this information to design modified immunogens that might selectively elicit “desirable”, broadly directed responses. For a more extensive description of our collaborators and a full list of publications relevant to this project, see our Flu Project web site.