A Computationally-Driven Predictive Framework For Stabilizing Viral Therapies

Project Personnel

Sarah Perry

Principal Investigator

University of Massachusetts, Amherst


Caryn Heldt

Michigan Technological University


Sapna Sarupria

Clemson University


Funding Divisions

Division of Materials Research (DMR), Division of Chemical, Bioengineering, Environmental, and Transport Systems (CBET)

Many vaccine production and delivery systems remain dependent on a cold chain requirement, which prevents millions of people from receiving vaccines annually. To increase the availability of current and future vaccines, the vaccine cold chain needs to be eliminated. While sugars and bulking agents are being explored to increase the thermal stability of viral vaccines, the cold chain is still the main method to stabilize viral vaccines. This is not only an issue for developing countries; proper temperature storage of vaccines is also a challenge in the US, with an outbreak of influenza having been potentially linked to improper vaccine refrigeration. A more standard and promising method to stabilize vaccine formulations is to add stabilizing excipients. With excipients, vaccines can be stored under refrigeration conditions. However, this approach has suffered from both a lack of generalizability and the absence of a fundamental understanding of the mechanism whereby stabilization is achieved. Empirical evidence has identified several excipients such as sugars, amino acids, and bulking agents like gelatin, dextran, and cellulose that help to stabilize proteins/viruses in both wet and dry formulations. In addition, it has been demonstrated that complex combinations of excipients (mixtures) are often used in final formulations. Experimental observations suggest that many of the excipients help to structure water and/or replace hydrogen-bonding interactions with the surface of the protein/virus to provide stability. However, most of the work published in this area has been empirical and experimental in nature and would be difficult to perform at the scale needed to elucidate the subtle ways in which molecular structure affects water structure and thus stability. 

In this project, a combination of experiments, modeling, and machine learning will be used to identify molecular features/motifs that impart this stability and use this framework to discover excipient mixtures for vaccine formulations. This approach has the potential to shift the paradigm for vaccine formulation – allowing for tailoring of formulations based on knowledge of the virus itself, rather than through an iterative, Edisonian process.