In April 2011, a group of researchers from the University of Washington described, in a paper published on Science, a computational method for designing proteins able to bind to the surface of a target macromolecule. They obtained a protein that binds a conserved surface patch on the stem of the influenza hemagglutinin from the 1918 H1N1 pandemic virus, thus inhibiting it. Recently, they manage to strongly improve the activity of this protein by applying an approach based on energy landscape, which is a mapping of all possible conformations of a molecular entity and their corresponding energy levels. Such a result is described in a paper appeared on the June issue of Nature Biotechnology.

The traditional approach to increase binding affinity in target molecules consists in multiple rounds of selection followed by conventional sequencing to identify the few best clones. The American researchers, however, choose a different strategy: they combined data from deep sequencing with the energy landscape mapping to optimize the search for the best binders. Then, they performed conventional directed evolution to select those binders. They combined large numbers of individually small, favorable effects that would have been very difficult to find by traditional affinity maturation approaches, thus obtaining a strong increase in binding affinity. The “upgraded” protein can bind all influenza group 1 hemagglutinins and neutralize H1N1 viruses as much effectively as many human antibodies do.

Thus, the combination of deep sequencing and computational protein design proved to be an effective tool to generate new therapeutic and diagnostic molecules with high affinity and specificity for their targets.