Marcus Noyes, PhD

In the Noyes Lab we have spent years studying how transcription factors determine what sequences of DNA they bind to and what genes they regulate.  Focusing on the two most common DNA-binding domains, the C2H2 zinc finger and the homeodomain, we have generated models of transcription factor specificity using only synthetically-derived data.  In fact, these data sample more protein space than all of the evolved DNA-binding domains combined, and by orders of magnitude!  As a result, our models outperform all other models for the prediction of homeodomain and zinc finger specificity. In addition to our understanding of transcription factor function, we have used our data to produce the first AI-based model of zinc finger design.  This model solved a 30-year design challenge for the zinc finger domain and made their ease of use on par with the CRISPR-Cas systems. This new model is incredibly impactful as it allows us to leverage the advantages of the zinc finger domain for therapeutics: size, delivery, immunogenicity.  We are currently exploiting this design tool for epigenetic editing where we develop novel therapies that address diseases associated with the mis-regulation of genes. As we expand our platform, we will apply our designer zinc finger domains as networks for cellular reprogramming and regenerative medicine.