Google announced it will soon release TxGemma, a collection of open AI models. Karen DeSalvo, chief health officer at Google, made the announcement at a recent Boston healthtech conference, according to TechCrunch; she also wrote about it in a blog post.
“The development of therapeutic drugs from concept to approved use is a long and expensive process, so we’re working with the wider research community to find new ways to make this development more efficient,” DeSalvo said.
She also said that Google will be making TxGemma available after this month, “to the community to build on and improve through Health AI Developer Foundations.”
According to Google, TxGemma understands regular text and the structures of different therapeutic entities, such as small molecules, chemicals, and proteins. Researchers can ask TxGemma questions to help predict the properties of potential new therapies.
This should fall into many eager hands. However, the models’ license terms are not yet clear.
There are an estimated several hundred AI startups in drug discovery, and the global market for AI in this field is expected to grow from $3.5 billion in 2023 to $7.9 billion by 2030, reflecting an annual growth rate of 12.2%.
Google was at the head of the field. When its DeepMind’s AlphaFold was unveiled in 2020, the new protein structure prediction program was hailed as a breakthrough that could address the high costs and dismal failure rates in drug discovery and development. AlphaFold uses an end-to-end deep neural network trained to produce protein structures from amino acid sequences, multiple sequence alignments, and homologous proteins.
Leveraging the success of AlphaFold, about two years after its launch, DeepMind’s AI spinout Isomorphic announced two drug discovery deals, worth $3 billion each, with Eli Lilly and Novartis.
In February of this year, Isomorphic announced it was expanding its strategic research collaboration with Novartis. Under the terms of the agreement, the two companies will expand the scope of the initial collaboration, adding up to three additional research programs on the same financial terms as the original agreement.
However, there have already been some high-profile failures using AI for drug discovery, including at Exscientia and BenevolentA. Yet, the momentum toward AI in drug discovery is big and getting bigger.
“The entire industry is pivoting to precision medicine,” Alban de La Sablière told Inside Precision Medicine in an earlier article. He is the chief operating officer of Owkin, which uses patient-data-based AI to speed clinical trials. “A lot of drugs have failed because companies were reaching for the sixth millionth mutation and it didn’t have that big an effect. Pushing an asset forward, but in a smaller subgroup using AI with the right data makes more sense.”
As Eric Topol, MD, wrote in a Substack post (October 1, 2023), “Not long ago, scientists might spend 2 or 3 years to define the 3-dimensional structure of a protein… Now that can be done for nearly all proteins in a matter of minutes, thanks to advances in AI. Even new proteins, not existing in nature, never previously conceived, can now be created.”
