Meet DARWIN AI, an artificial intelligence algorithm that powers the NuvoBio platform. Using DARWIN AI, NuvoBio is able to design very small proteins (i.e., peptides) with therapeutic interest.

To date, NuvoBio has successfully developed and experimentally validated:

  • In vitro and Cell-active enzyme inhibitor peptides
  • Extracellular receptor binding peptides
  • Immuno-regulatory peptides
  • Anti-microbial peptides
  • Beta-lactamase inhibitor peptides, and
  • Anti-viral peptides

What makes DARWIN AI unique?

Unlike many other computationally-intensive approaches to peptide discovery, DARWIN AI does not rely on a known crystal structure, and instead, relies on learning how proteins interact naturally and evolving novel peptides based on these principles. The result is a unique AI-based approach toward peptide inhibitor development that has been validated experimentally.

​A key advantage of DARWIN AI is its independence from known structural information. This additional freedom allows DARWIN AI to be implemented against targets that arise from new genomes (i.e., novel viruses) or uncharacterized targets with only primary sequence and interaction information available.

Watch Darwin AI

DARWIN AI works by evolving peptide(s) of interest (e.g., binding peptide(s)) over multiple generations, striving towards its ultimate goal of maintaining target interaction while eliminating interaction with non-targets (i.e., off-target effects).

Here, you will see an example of DARWIN AI evolving a therapeutic peptide that is specific to the NPR guanylate cyclase receptor (NPR1-3).

DARWIN AI is evolving binding peptides that are specific to NPR1 (and NPR2-3 family members) and strengthening this specificity within the Human proteome (approximately 20,000 putative human off-targets).

The result is a new peptide that is converged towards the FDA-approved peptide, Nesiritide. Nesiritide is a highly specific peptide for Natriuretic Peptide Receptor (NPR1-3) with established clinical implications.

Case study examples that showcase the ability of DARWIN AI to identify FDA-approved peptides:


Case Study 1: Can DARWIN independently confirm targets of FDA-approved therapeutic peptides?

Case Study 2: Can DARWIN converge on the peptide sequence of an FDA-approved therapeutic peptide?