The cell and gene therapy field is, some say, at a tipping point—poised to take off. The FDA has noted that, by next year, they aim to approve 10–20 new cell and gene therapies per year. To do that, however, multiple hurdles still need clearing. Or, at least, bypassing.

How does the gene therapy field intend to advance so quickly? Enter artificial intelligence (AI). For example, the scientists at Form Bio are building AI tools that, they say, provide an in silico solution to biotech companies and pharma companies working in gene therapy—with a focus on the genetic payload.

In order to build its proprietary large language model (LLM) and AI model for nucleic acids, Form Bio analyzed over three trillion nucleotides, explained Kent Wakeford, JD, co-CEO and co-founder of Form Bio. The model analyzes different components: DNA quality, expression, truncation, CpG islands, and mRNA concentration rates.

Now, the company announces a partnership with Ginkgo Bioworks to bring their AI out of the computer and into production. The two companies are currently teaming up on gene therapy, however they may have broader capabilities across nucleic acids, delivery platforms, and therapeutic modalities in the future.

From woolly mammoths to drugs

Form Bio is not the only company Wakeford has founded recently. He was also part of the group that founded Colossal Biosciences—the company focused on resurrection genomics projects such as bringing back the wooly mammoth. While at Colossal, Wakeford says, he realized that the tools they were building to help address some of the challenges with loss of biodiversity and species extinction could be relevant to improving human health. In 2022, Wakeford left Colossal to start Form Bio—which has offices in South San Francisco and Dallas, with other team members in Austin, Texas at the University of Texas Southwestern, Seattle, and elsewhere. Thirty Colossal employees made the move with Wakeford, and the addition of about 20 other scientists (including many from David Baker’s lab at the University of Washington) has built the 50-person strong team at Form Bio.

Tackling gene therapies hurdles

Form Bio’s AI plans to tackle gene therapy hurdles by optimizing the genetic payload. More specifically, by predicting truncated genomes, the distribution of genomes in AAVs, potential immunotoxic events, and flagging of problematic motifs—like secondary or tertiary structures that lead to problems in manufacturing. This work, Wakeford asserts, could reduce the number of empty capsids, result in higher capsid titers, correct protein expression, lead to more functional proteins, and lower immunogenicity.

Form Bio
The Form Bio leadership team includes (from left) Andrew Busey, co-CEO; Brandi Cantarel, PhD, director of bioinformatics; Kent Wakeford, co-CEO; Claire Aldridge, PhD, chief strategy officer; and Ben Lamm, co-founder. [Photo: John Davidson/Form Bio]
They can, they say, analyze an initial drug and predict what is going to happen along the way.  “We’ll tell you where it is going to break down in manufacturing and where you’re going to have issues within your genome. We’re going to tell you a roadmap of the problems that you’re going to have that you should address in preclinical development.”

It also moves the process along more quickly. Wakeford notes that they can speed things up to a time that is “a fraction of what you normally see in drug development.”

From the computer to the wet lab

Ginkgo Bioworks has significantly expanded its gene therapy capabilities in recent years. For example, it acquired Circularis’ proprietary circular RNA and promoter screening platform, partnered with Biogen to improve AAV manufacturing, and partnered with Selecta Biosciences to develop AAV capsids with altered tropism and immunogenicity.

To that end, its existing suite of capabilities for gene therapy include licensable capsid assets, capsid engineering services, payload engineering and manufacturing optimization services.

Form Bio may be able to predict what will happen with a gene therapy, but Ginkgo has the capabilities to test those predictions.

“Ginkgo’s gene therapy platform is focused on capsid engineering, payload engineering, and AAV production. Our AI optimized designs fit right into their engineering and production. Whether just their engineering work, or their existing capsid and/or AAV production host cell lines,” notes Wakeford.

That said, Ginkgo does not make drugs; it is not a contract development and manufacturing organization (CDMO). A biotech company working with Form Bio will plan to take what it learns from Ginkgo’s platform to a bioprocessing facility in order to produce the drug.

Form Bio has not released data yet on their platform (though they assert that there will be data coming in the near future.) And the promise of the application of AI to drug discovery elicits both excitement and raised eyebrows. As for Wakeford, he is “giddy” about the promise of AI and gene therapy, and where Form Bio may fit in.

Previous articleBrain Tumors, Infections May Have an Achilles’ Heel … in the Eye
Next articleBiotech around the World: The Low-Down on the Land Down Under