Research Project
10101230
An immense advancement in machine learning and artificial intelligence has transformed many aspects of our lives. The integration of artificial intelligence into the biomedical field allows us to solve complex biological problems that are the bottle neck of developing progressive diagnostic and therapeutic tools. One example is the need to manipulate the amino acid sequence of peptides to improve their function as bioactive molecules. Embarking on these new technologies, we developed a new machine learning tool that is based on a discipline known as ?genetic programing? that can assist in designing new proteins and bioactive peptides. This new technology, termed Protein Optimization Evolving Tool (POET), can generate a model that describes the relationship between a peptide and its respective activity. Moreover, through cycles of protein evolution, we can significantly improve the model and consequently generate peptides with substantially improved function. A major challenge of translating synthetic biology approaches to clinical treatment is the need to improve the communication with biological circuits in vivo. To that end, we will leverage the immense potential of the POET to produce proteins and peptides that can read and write information from and into cells. Here we seek to improve, test and implement this model into three related, yet, independent aims. In the first aim, we will deploy the POET to develop an ultrasensitive peptide-based imaging agent for MRI based on proton exchange. Our preliminary data shows that through only few cycles of peptide evolution we surpassed the state-of-the-art similar peptides. In the second aim, we intend to use a similar approach to develop a novel MRI imaging probe based on T1 relaxation. We will use a metabolic engineering approach to express and load the peptide with Lanthanides, and the POET algorithm to improve the next generations. Lastly, in the third aim, we will use the POET for discovering new peptides for drug and gene delivery. We will utilize a novel platform for gene/drug delivery to test the efficiency of the peptides. All three aims will start with computational design of peptides followed by an in vitro testing and several cycles of peptide evolution until the ultimate peptides are identified. All three aims will be ended by demonstration of the utility of those peptides in a clinically relevant question in an in vivo model followed by non-invasive imaging. We anticipate that this innovative approach will open up a new avenue for developing powerful bioactive peptides and proteins to solve critical biological questions, and for developing new diagnostic and therapeutic approaches that can vastly benefit the well-being of numerous patients.
