Artificial Intelligence Creates AAV Capsid for Gene Therapy

2019-12-01

Dyno Therapeutics, a groundbreaking biotechnology company that applies artificial intelligence to gene therapy, has published an article in the journal Science today, showcasing the power of a comprehensive machine-guided approach that can provide improved capsids for gene therapy. This research was led by Dr. Eric D. Kelsic and Dr. Sa M. Sinai, co-founders of Dyno and researchers at the Wyss Institute for Biologically Inspired Engineering at Harvard. The publication is titled "Comprehensive AAV Capsid Adaptation Landscape Reveals Viral Genes and Enables Machine-Guided Design". AAV capsids are currently the most commonly used gene therapy vectors as they can deliver genetic material to patient organs and have demonstrated safety. However, naturally occurring AAV capsids lack key features for optimal gene therapy, such as targeted delivery, immune system evasion, higher levels of virus production, and increased transduction efficiency. Since 2015, researchers have been working to overcome the limitations of current capsids by developing new machine-guided techniques to rapidly and systematically design a new set of improved capsids for broad therapeutic applications. In their study published in Science, the authors demonstrate the progress of their unique machine-guided AAV engineering method. Previous approaches to engineering capsids have been challenging as they often disrupt the functionality of the capsid proteins and lack knowledge about how AAV capsids interact with the body. Historically, the most popular capsid engineering methods have not directly addressed this issue and instead employed an indirect solution: generating new capsids by randomly altering proteins. However, most random changes to capsids actually result in a loss of function, leading to a scarcity of viable capsids in these random libraries and even fewer improved capsids. Recognizing the limitations of traditionally generated Capid libraries, the authors implemented a machine-guided approach, using new high-throughput measurement techniques to gather extensive data and teach them how to construct better libraries, ultimately resulting in synthetic Capids with optimized delivery properties. Focusing on the AAV2 capsid, the authors constructed a complete library of single codon substitutions, insertions, and measured their in vivo delivery characteristics. They then effectively generated multiple diverse AAV capsid libraries for mouse liver targeting using the machine-guided approach, which outperformed AAVs generated by traditional random mutagenesis methods. Throughout this process, the authors revealed previously unidentified proteins encoded within all the most popular AAV capsid sequences, known as membrane-associated accessory proteins. These proteins are believed to play a significant role in the natural lifecycle of AAV. Dyno Therapeutics, founded by experienced biotech entrepreneurs and leading scientists in synthetic biology, gene therapy, and machine learning, is revolutionizing gene therapy by applying artificial intelligence. The improved version of AAV capsids will redefine this therapy.

RECENT AI NEWS

RECENT AI TOOLS