President Trump has tested positive for coronavirus

President Trump, who has frequently dismissed the significance of the Covid-19 pandemic and rarely wears masks in public, has contracted the coronavirus and is now in quarantine, he announced early Friday morning on Twitter. Melania Trump has also tested positive for the SARS-CoV-2 virus, the president said. The news came a few hours after a number of news outlets revealed that Hope Hicks, a top adviser to the president, had contracted the virus. Hicks, who regularly travels with Trump, was part of his preparatory team for Tuesday’s debate against Democratic rival Joe Biden, STAT's Helen Branswell writes.

Within minutes of Trump’s announcement, disease experts began questioning whether there was a larger outbreak among White House staff, pointing to the fact it would be unlikely the president would test positive so quickly, if Hicks had transmitted the virus to him and his wife.

Read more.

Congress puts more drug execs on the spot

Executives from Amgen, Novartis, and Mallinckrodt were forced to explain their drug cost hikes to Congress Thursday as part of an 18-month investigation into pricing practices and compensation packages. The line of questioning was largely tedious, as it was Wednesday when Congressional Democrats grilled execs from Celgene, Bristol Myers Squibb, and Teva Pharmaceuticals. But there were some shining moments of awkwardness, STAT’s Nicholas Florko writes.

For example, one tense moment sprang up when Rep. Katie Porter (D-Calif.) pulled out a whiteboard and asked Amgen CEO Robert Bradway to list the reasons why he and five colleagues deserved to earn $124 million over the course of three years. When Bradway couldn’t provide a satisfactory answer, Porter wrote on the white board: “????”

Does an itchy arm ruin a clinical trial?

Does intermittent fasting work? And is pharma more trustworthy than the CDC? We discuss all that and more this week on “The Readout LOUD,” STAT’s biotech podcast. First, Yale University vaccine expert Saad Omer joins us to discuss how the side effects of Covid-19 vaccines could disrupt ongoing clinical trials.

Next, University of California, San Francisco, cardiologist Ethan Weiss calls in to tell us about new clinical trial results that call into question the benefits of intermittent fasting. Finally, we bring you a lightning round, with hot takes on the first presidential debate, new data on a Covid-19 treatment, and a spicy congressional hearing.

Listen here.

Pulling out the ol’ Nobel crystal ball

The winners of this year’s Nobels will be unveiled next week, and there are always obvious-seeming contenders for the big prize. But as MIT’s Phillip Sharp, a 1993 Nobel-winner, told STAT’s Sharon Begley: “There is just a lot of good science that will never get recognized.” So instead of focusing on biology’s greatest hits, like, say, CRISPR gene editing, why not scope out some outliers that are still scientific standouts?

There are plenty of biological possibilities that might be worthy of a Nobel, a trove of experts tell STAT. And there seems to be something of a pattern in how these prizes are chosen. For instance, in the past 20 years of Nobel Prizes in medicine and physiology, 83% had won another major prize beforehand: the Lasker, the Gairdner, or the Horwitz prize, one Nobel prediction aficionado said.

Perlmutter unpacks AI in drug discovery

Roger Perlmutter, the head of research and development at Merck, is joining the board of Insitro, a firm focused on using artificial intelligence to discover drugs. He got on Zoom to talk with STAT’s Matt Herper about the potential of AI in drug discovery.

“The hardest part of drug discovery and development — it’s all hard, it’s all hard — but the hardest part is picking targets,” Perlmutter said. “That’s where it all starts. And picking targets is at this point is largely an intuitive process. We talk about scientists who have a good taste. Well, you know, I want scientists to have good taste, too, but you can’t rely on it.”

Could AI help? It's possible machine learning could be "very effective if you can get a large enough data set," Perlmutter said.