A deep dive into the meaning of AI drug development

Biotech companies like Absci and Generate:Biomedicines have raised hundreds of millions of dollars by claiming to design novel antibody drugs from scratch using artificial intelligence. However, a STAT examination of company documents and expert interviews shows that these companies routinely exaggerating their capabilities.

Instead of true, de novo drug discovery, they primarily optimize existing antibodies — a process that has been possible for decades without AI, reports STAT's Brittany Trang.

These companies defend their approach, saying they're achievements are stepping stones toward fully AI-generated drugs. Critics argue that their marketing misleads investors and the public.

"This is stuff of already a few decades ago — not that it's not useful; this is good stuff. But it's not AI," one antibody researcher told STAT.

RESEARCH

Will Trump cost-cutting endanger genetic databases?

Scientists are increasingly concerned that the Trump administration's cost-cutting — and ideological bent — may threaten vital genetic databases like GenBank and the Sequence Read Archive, which are important tools in global biomedical research. These repositories, which are maintained by NIH's National Library of Medicine, store invaluable genetic sequence data used to study diseases, track pandemics, and develop treatments.

While there has been no explicit threat to the databases, experts worry that political interference, budget cuts, or the administration's Silicon Valley-style "move fast and break things" approach could compromise their integrity, slowing scientific progress. And hobbling these technologies is something of an existential threat: GenBank contains the genetic information about every single living thing that has had its DNA sequenced.

"I think we need all this technology to sustain the density of life on earth," one academic who did not want to be named told STAT's Helen Branswell.

trump administration

NIH abruptly cuts diversity grants

The NIH has halted the allowance of F-31 diversity grants and other funding programs that support researchers from marginalized backgrounds, complying with President Trump's executive orders to ban DEI initiatives. The decision has sparked widespread concern among some scientists, as these grants have been crucial in helping underrepresented Ph.D. students, postdocs, and faculty establish scientific careers.

Critics argue this move will exacerbate the existing disparities in science funding, and universities and private institutions will be left to fill the gap.

"Why is this important? If we limit the people who can do biomedical research then we are not supporting a society that's becoming — whether we like or not — more diverse," a University of Kansas economist studying race and gender in science told STAT's Usha Lee McFarling and Anil Oza. "And when we limit research by certain people, or on certain topics, we are leaving discoveries on the table."

antibiotics

FDA needs to reconsider its approach to antibiotics

The FDA's approach to antibiotic approvals focuses too much on killing bacteria in test tubes rather than improving patient outcomes, opines Diana Zuckerman, president of the National Center for Health Research, and John H. Powers III, a clinical medicine professor at George Washington University.

Antibiotic resistance is a concern, to be sure, but most Americans who die from infections do so despite the availability of FDA-approved treatment options: Regulators have approved a dozen new antibiotics that can kill MRSA in a lab setting, they say, but they haven't been shown to improve lives of patients with these infections.

The issue is that the new drugs targeting resistant bacteria aren't actually given to the majority of patients. And many of the newly approved antibiotics haven't been proven to save more lives than existing treatments, they point out.

More around STAT