'Google promises new AI models for drug discovery'

 Google gave an update on various aspects of its health-related R&D at its annual event The Check Up, including a new collection of artificial intelligence (AI) models intended to support drug discovery.

The TxGemma models are – as the name suggests – an offshoot of Google's Gemma family of open-source, generative AI (GenAI) models that are, in turn, based on its Gemini AI platform, the latest iteration of which launched in December.

The toolkit is due to be made available to the scientific community for appraisal and further development later this month through its Health AI Developer Foundations programme, although it's not clear yet whether they can be used and adapted for commercial applications.

According to Google's chief health officer Dr Karen DeSalvo, TxGemma is able to understand regular text and the structures of different therapeutic entities, such as small molecules, chemicals and proteins.

"This means researchers can ask TxGemma questions to help predict important properties of potential new therapies, like how safe or effective they might be," she wrote 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," said DeSalvo.

What is clear is that the use of AI has emerged as a transformative force in the life sciences industry thanks to its ability to process large-scale datasets, uncover patterns, and generate predictions, and it is already being used to find drug targets, design new drugs, and repurpose existing therapies, among other applications.

Its rapid adoption prompted the FDA to publish its first guidance on the technology's use in regulatory filings earlier this year, while, in 2024, the EMA also set out its thinking in a reflection paper that discussed how AI can be used in the medicinal product lifecycle.

At The Check Up, Google also provided a rundown of its progress in other health applications, including improved health results in Google search, a new medical records feature in its Health Connect app, and its AI 'co-scientist' – a virtual collaborator designed to help scientists generate novel hypotheses and research proposals – that was announced in February.

"Let's say researchers want to better understand the spread of a disease-causing microbe," said DeSalvo. "They can specify this research goal using natural language, and the AI co-scientist will propose testable hypotheses, including a summary of relevant published literature and a possible experimental approach."

Finally, Google also highlighted an AI tool called Capricorn that uses Gemini models to help physicians speed up the identification of suitable personalised treatments for childhood cancers by combining public medical and de-identified patient data.

https://pharmaphorum.com/news/google-promises-new-ai-models-drug-discovery