Cutting-edge, machine-learning techniques are increasingly being adapted and applied to biological data, including for COVID-19.

The big picture: Discovering and developing a new drug typically takes more than a decade and costs on average close to $1 billion, making it difficult to build a cache of pharmaceuticals to fight future pandemics or stop intractable diseases.

• But scientists are combining two scientific leaps — in machine-learning algorithms and powerful biological imaging and sequencing tools — to try to spur progress in understanding diseases and advance AI itself.

What's happening: Last month, researchers reported using a new technique to figure out how genes are expressed in individual cells and how those cells interact in people who had died with Alzheimer's disease, allowing scientists to better understand the development of a condition that afflicts nearly 6 million Americans.

• Machine-learning algorithms can also be used to compare the expression of genes in cells infected with SARS-CoV-2 to cells treated with thousands of different drugs in order to try to computationally predict drugs that might inhibit the virus, says Kris White, who studies antiviral treatments at the Icahn School of Medicine at Mount Sinai and is conducting research that uses this approach and has not yet been published.
• Yes, but: Algorithmic results alone don't prove the drugs are potent enough to be clinically effective. But they can help identify future targets for antivirals or they could reveal a protein researchers didn't know was important for SARS-CoV-2, providing new insight on the biology of the virus, says White.

These and other advances are fueling centers and startups that apply AI to drug discovery, including a $400 million investment last month in insitro, which uses machine learning and genomics data to identify new drug candidates or existing ones that can be repurposed for treating neurodegenerative diseases.

Between the lines: AI's power lies in its ability to sift through data and find useful patterns far more quickly than humans could do alone. So far that's mostly been used to turbocharge movie, song and product recommendations on the web.

• But "biology requires more than that," says Caroline Uhler, the co-director of the Eric and Wendy Schmidt Center, a $300 million initiative launched last month at the Broad Institute to focus on the intersection of machine learning and biology.
• For example, creating a drug to target a protein involved in a disease requires understanding how the genes that give rise to that protein are regulated.

One of the challenges is that biological data comes in many forms: DNA code of course, but also the expression of genes, electrical signals, images and more that are all captured at different snapshots in time and then have to be stitched together to get a full picture of what is happening in a patient.

What to watch: As with AI's application in other fields, there are real concerns about bias in datasets used to build machine-learning models.

• An overreliance on genomic data from people of a particular race, gender or geography can skew an algorithm's predictions for other people, raising concerns about misdiagnoses or unequal health care.

The intrigue: Much of the focus is on what AI can do for medicine and biology, but Uhler and her Broad Institute co-director Anthony Philippakis — who both have backgrounds in statistics — say the nuanced, complex biological data captured by imaging and sequencing could help to create powerful algorithms that capture cause and effect in a system.

• That would represent a leap forward for AI, which remains best at identifying correlations, while leaving the question of cause to human scientists.

The bottom line: Melding AI and biology may not just be another tool for understanding medicine. Biology could be "a driver for the next generation of advances in machine learning," says Philippakis.

Go deeper.

Antiviral drugs can be a key pandemic-fighting tool, but so far there's only one approved in the U.S. for COVID-19.

Why it matters: Because some people won't get vaccinated, and because there will likely be new variants of the virus, we'll need effective treatments — including antivirals, former FDA commissioners Scott Gottlieb and Mark McClellan wrote earlier this week in the Wall Street Journal.

Driving the news: This week, Merck and Ridgeback Biotherapeutics said the antiviral molnupiravir "significantly reduced infectious virus in subjects in a mid-stage study after five days of treatment," WSJ reported.

• Merck is expecting interim results from two later-stage clinical trials looking at whether the drug prevents hospitalization and death from the virus. Those studies will help determine whether the drug has a clinical benefit.
• On Tuesday, Pfizer presented details about an oral antiviral it developed from scratch during the pandemic, Chemical & Engineering News reports.

How it works: Antivirals stop a virus from reproducing — either as it attaches to a person's cells, uses those cells to make copies of itself, or exits the cells for the rest of the body.

• Some antivirals target proteins in the virus itself. Others inhibit machinery in the host cell that the pathogen relies on to replicate.
• The catch is that all have to be given early in the course of the disease to prevent the virus from getting a foothold and spreading in the body.

Where it stands: So far, remdesivir — a drug investigated earlier to treat Ebola and other diseases — is the only antiviral approved in the U.S. for COVID-19.

• While there is evidence it helps speed recovery and prevents the disease from progressing, it doesn't prevent death from the virus. And because it is administered via an IV, the drug's use is limited to people who have been hospitalized — and are therefore further into their illness.
• 31 other antivirals are being investigated, according to the Milken Institute's COVID-19 tracker.
• A goal for SARS-CoV-2 is to develop an equivalent to Tamiflu, an oral antiviral that can be taken at home after someone is exposed but before symptoms appear, says Armand Balboni, CEO of Appili Therapeutics, which is studying the effectiveness of another antiviral — favipiravir — for COVID-19.

What to watch: There are concerns that current vaccines and antibody treatments could be less effective against emerging SARS-CoV-2 variants because they act on proteins on the surface of the virus that can mutate quickly.

• Antivirals target key machinery for virus replication that tends to be more stable, but resistance to antivirals can develop too, White says.
• An antiviral that targets the host cell's proteins versus the virus' proteins is one approach to try to minimize resistance. Combining antivirals, similar to what is done to treat HIV, is another avenue under investigation for COVID-19.

Go deeper.

"The U.S. averaged about 65,000 new cases per day over the past week, essentially unchanged from the week before," per Axios' Sam Baker and Andrew Witherspoon.

The B.1.1.7 coronavirus variant, first detected in the U.K., is now the dominant strain in the U.S., the CDC said Wednesday, Axios' Shawna Chen reports. Check out Axios' Coronavirus Variant Tracker.

The CDC this week acknowledged the risk of getting COVID-19 from a contaminated surface is low, the NYT's Emily Anthes reports.

Most kids with MIS-C (multisystem inflammatory syndrome in children) reported few or no COVID symptoms in a study of 1,075 children with the condition, Kerry Grens writes in The Scientist.

The results of high-energy physics experiments released on Wednesday open the possibility that a tiny subatomic particle called a muon may act in ways that break the known laws of physics, Axios' Bryan Walsh writes.

The big picture: The experimental work — while still far from conclusive — underscores the fact that science still has much to learn about the fundamental workings of the universe, and it points the way toward further breakthroughs.

Driving the news: In a news conference and virtual seminar on Wednesday, as well as a set of papers published the same day, scientists announced the first results of the Muon g-2 experiments being carried out at the Fermi National Accelerator Laboratory, or Fermilab.

• Muons are subatomic particles similar to electrons but possess 207 times as much mass — hence the rather unflattering nickname "fat electrons."
• The particles — which have puzzled scientists since they were first discovered in 1936 — are produced in large amounts during collider experiments at places like Fermilab that involve smashing particles together at high speeds.

What they found: When the muons were sent through intense magnetic fields at Fermilab's Muon g-2 ring, they behaved in ways that didn't quite line up with theoretical predictions, wobbling more than expected.

• Anytime nature throws us a curveball, scientists take notice, and the fact that the Fermilab experiments lined up with similar work at Brookhaven National Laboratory in 2001, which has long puzzled researchers, is notable.
• The experiments suggest the Standard Model — physics' fundamental theory about how particles interact with each other — may be far from complete.

The catch: The scientists behind the experiments reported that the results had a 1 in 40,000 chance of being a fluke — pretty good, but still short of the certainty required to claim an official discovery in physics.

You probably have an asymptomatic infection right now (Sarah Zhang — The Atlantic)

Studies that make brainlike structures or add human cells to animal brains are ethical, panel says (Jocelyn Kaiser — Science)

NASA's InSight lander feels Mars quake beneath it (Miriam Kramer — Axios)

Scouring the desert for a seabird (Rachel Fritts — Hakai)

Let's hope the coming cicadas are this cute.

After Hurricane Maria devastated the island home of about 1,500 rhesus macaques, the monkeys coped by extending their social networks to make new friends, according to new research.

The big picture: The findings speak to just how socially flexible animals, including humans, can be and the key role of not just our closest allies but also others in enduring or surviving extreme events.

• "That is maybe one big advantage we have as gregarious primates — to use our social relationships for our advantage," says Camille Testard, lead author of the paper published today in Current Biology and a graduate student at the University of Pennsylvania.

What they found: In the days and weeks after the hurricane in 2017 decimated Cayo Santiago in Puerto Rico, researchers noticed the macaques, which can be quite competitive, actually seemed to be less aggressive and more tolerant of one another.

• Testard and her colleagues tapped a trove of data about the monkeys' behavior on the long-established research site and looked at two bonding behaviors — sitting next to one another and grooming one another — three years before and one year after the hurricane.
• Analyzing their social networks, the researchers found that instead of spending more time strengthening their relationships with existing partners, the monkeys invested time in developing new relationships and bonds. (They also maintained relationships with those existing key partners.)
• That was especially true among macaques that had been less sociable before the hurricane, they report.
• Driven most likely by a need to find shade after the storm destroyed more than half the island's vegetation, the macaques may have made nice with other macaques — oftentimes friends of friends or neighbors — in case they needed to join them for refuge from the heat, Testard says.

What's next: Testard wants to look at how long the changes lasted, the biological differences between the animals that made lots of new connections and those that didn't, and the long-term consequences of being able to make new friends.