Short film is alive and well. Using the current trendy gene-editing system CRISPR, a team from Harvard University has encoded images and a short movie into the DNA of living bacteria.

The work is part of a larger effort to use DNA to store data — from audio recordings and poetry to entire books on synthetic biology. Last year, Seth Shipman and his colleagues at Harvard threw CRISPR into the mix when they used the editing system to record molecular data in the DNA of Escherichia coli.

Now, the team is upping its game with images of a human hand and a short movie, a GIF of a galloping horse from iconic turn-of-the-century photographer Eadweard Muybridge’s Human and Animal Locomotion. In the code, the nucleotide bases that form DNA correspond to black-and-white pixel values. The video was encoded frame by frame. Once the team synthesized the DNA, they used CRISPR and two associated Cas proteins (Cas 1 and 2) to slip the data into the genetic blueprint of E. coli colonies.

After growing the bacteria for several generations, the scientists retrieved the code for the images and film frames and were able to reconstruct the clips. About 90 percent of the encoded information was left intact. Though it’s not a perfect storage system, the results demonstrate CRISPR’s potential for hiding data in the genetic blueprints of bacteria, Shipman and his colleagues write July 12 in Nature.

When my younger daughter was around 6 months old, we gave her mashed up prune. She grimaced and shivered a little, appearing to be absolutely disgusted. But then she grunted and reached for more.

Most babies are ready for solid food around 6 months of age, and feeding them can be fun. One of the more entertaining approaches does not involve a spoon. Called baby-led weaning, it involves allowing babies to feed themselves appropriate foods.

Proponents of the approach say that babies become more skilled eaters when allowed to explore on their own. They’re in charge of getting food into their own mouths, gumming it and swallowing it down — all skills that require muscle coordination. When the right foods are provided (yes to soft steamed broccoli; no to whole grapes), babies who feed themselves are no more likely to choke than their spoon-fed peers.

Some baby-led weaning proponents also suspected that the method might ward off obesity, and a small study suggested as much. The idea is that babies allowed to feed themselves might better learn how to regulate their food intake, letting hunger and fullness guide them to a reasonable calorie count. But a new study that looked at the BMIs of babies who fed themselves and those who didn’t found that babies grew similarly with either eating style.

A clinical trial of about 200 mother-baby pairs in New Zealand tracked two different approaches to eating and their impact on weight. Half of the moms were instructed to feed their babies as they normally would, which for most meant spoon-feeding their babies purees, at least early on. The other half was instructed that only breast milk or formula was best until 6 months of age, and after that, babies could be encouraged to feed themselves. These mothers also received breastfeeding support.

At the 1- and 2-year marks, the babies’ average BMI z-scores were similar, regardless of feeding method, researchers report July 10 in JAMA Pediatrics. (A BMI z-score takes age and sex into account.) And baby-led weaning actually produced slightly more overweight babies than the other approaches, but not enough to be meaningful. At age 2, 10.3 percent of baby-led weaning babies were considered overweight and 6.4 percent of traditionally-fed babies were overweight. The two groups of babies seemed to take in about the same energy from food, analyses of the nutritional value and amount of food eaten revealed.

The trial found a few other differences between the two groups. Babies who did baby-led weaning exclusively breastfed for longer, a median of about 22 weeks. Babies in the other group were exclusively breastfed for a median of about 17 weeks. Babies in the baby-led weaning group were also more likely to have held off on solid food until 6 months of age.

While baby-led weaning may not protect babies against being overweight, the study did uncover a few perks of the approach. Parents reported that babies who fed themselves seemed less fussy about foods. These babies also reportedly enjoyed eating more (though my daughter’s prune fake-out face is evidence that babies’ inner opinions can be hard to read). Even so, these data seem to point toward a more positive experience all around when using the baby-led weaning approach. That’s ideal for both experience-hungry babies and the parents who get to savor watching them eat.

Type 2 diabetes and prion disease seem like an odd couple, but they have something in common: clumps of misfolded, damaging proteins.

Now new research finds that a dose of corrupted pancreas proteins induces normal ones to misfold and clump. This raises the possibility that, like prion disease, type 2 diabetes could be triggered by these deformed proteins spreading between cells or even individuals, the researchers say.

When the deformed pancreas proteins were injected into mice without type 2 diabetes, the animals developed symptoms of the disease, including overly high blood sugar levels, the researchers report online August 1 in the Journal of Experimental Medicine.
“It is interesting, albeit not super-surprising” that the deformed proteins could jump-start the process in other mice, says Bruce Verchere, a diabetes researcher at the University of British Columbia in Vancouver. But “before you could say anything about transmissibility of type 2 diabetes, there’s a lot more that needs to be done.”

Beta cells in the pancreas make the glucose-regulating hormone insulin. The cells also produce a hormone called islet amyloid polypeptide, or IAPP. This protein can clump together and damage cells, although how it first goes bad is not clear. The vast majority of people with type 2 diabetes accumulate deposits of misfolded IAPP in the pancreas, and the clumps are implicated in the death of beta cells.

Deposits of misfolded proteins are a hallmark of such neurodegenerative diseases as Alzheimer’s and Parkinson’s as well as prion disorders like Creutzfeldt-Jakob disease (SN: 10/17/15, p. 12).

Since IAPP misfolds like a prion protein, neurologist Claudio Soto of the University of Texas Health Science Center at Houston and his colleagues wondered if type 2 diabetes could be transmitted between cells, or even between individuals. With this paper, his group “just wanted to put on the table” this possibility.

The mouse version of the IAPP protein cannot clump — and mice don’t develop type 2 diabetes, a sign that the accumulation of IAPP is important in the development of the disease, says Soto. To study the disease in mice, the animals need to be engineered to produce a human version of IAPP. When pancreas cells containing clumps of misfolded IAPP, taken from an engineered diabetic mouse, were mixed in a dish of healthy human pancreas cells, it triggered the clumping of IAPP in the human cells.
The same was true when non-diabetic mice got a shot made with the diabetic mouse pancreas cells. The non-diabetic mice developed deposits of clumped IAPP that grew over time, and the majority of beta cells died. When the mice were alive, more than 70 percent of the animals had blood sugar levels beyond the healthy range.

Soto’s group plans to study if IAPP could be transmitted in a real world scenario, such as through a blood transfusion. They’ve already begun work on transfusing blood from mice with diabetes to healthy mice, to see if they can induce the disease. “More work needs to be done to see if this ever operates in real life,” Soto says.

Even if transmission of the misfolded protein occurs only within an individual, “this opens up a lot of opportunities for intervention,” Soto says, “because now you can target the IAPP.”

Verchere also believes IAPP is “a big player” in the progression of type 2 diabetes, and that therapies that prevent the clumps of proteins from forming are needed. Whether or not future research supports the idea that the disease is transmissible, the study is “good for appreciating the potential role of IAPP in diabetes.”

Cross two flashlight beams and they pass right through one another. That’s because particles of light, or photons, are mostly antisocial — they don’t interact with each other. But now scientists have spotted evidence of photons bouncing off other photons at the Large Hadron Collider at CERN, the European particle physics lab in Geneva.

“This is a very basic process. It’s never been observed before, and here it is finally emerging from the data,” says theoretical physicist John Ellis of King’s College London who was not involved with the study. Researchers with the ATLAS experiment at the LHC report the result August 14 in Nature Physics.
Because photons have no electric charge, they shouldn’t notice one another’s presence. But there’s an exception to that rule. According to quantum mechanics, photons can briefly transform into transient pairs of electrically charged particles and antiparticles — such as an electron and a positron — before reverting back to photons. Predictions made more than 80 years ago suggest that this phenomenon allows photons to interact and ricochet away from one another.

This light-by-light scattering is extremely rare, making it difficult to measure. But photons with more energy interact more often, providing additional chances to spot the scattering. To produce such energetic photons, scientists slammed beams of lead nuclei together in the LHC. Photons flit in and out of existence in the lead nuclei’s strong electromagnetic fields. When two nuclei got close enough that their electromagnetic fields overlapped, two photons could interact with one another and be scattered away.

To measure the interaction, ATLAS scientists sifted through their data to find collisions in which only two photons — the two that scattered away from the collision — appeared in the aftermath. “That’s the trickiest part of the whole thing,” says physicist Peter Steinberg of Brookhaven National Laboratory in Upton, N.Y., a member of the ATLAS collaboration. The scientists had to ensure that, in their enormous, highly sensitive particle detector, only two photons appeared, and convince themselves that no other particles had gone unaccounted for. The researchers found 13 such events over 19 days of data collection. Although other processes can mimic light-by-light scattering, the researchers predict that only a few such events were included in the sample.

The number of scattering events the researchers found agrees with the predictions of the standard model, physicists’ theory of particle physics. But a more precise measurement of the interaction might differ from expectations. If it does, that could hint at the existence of new, undiscovered particles.

Spacecraft have gone bite-sized. On June 23, Breakthrough Starshot, an initiative to send spacecraft to another star system, launched half a dozen probes called Sprites to test how their electronics fare in outer space. Each Sprite, built on a single circuit board, is a prototype of the tiny spacecraft that Starshot scientists intend to send to Alpha Centauri, the trio of stars closest to the sun. Those far-flung probes would be the smallest working spacecraft yet.

“We’re talking about launching things that are a thousand times lighter than any previous spacecraft,” says Avi Loeb, an astrophysicist at Harvard University who is part of the committee advising the initiative. A Sprite is only 3.5 centimeters square and weighs four grams, but packs a solar panel, radio, thermometer, magnetometer for compass capabilities and gyroscope for sensing rotation.

These spacecraft are designed to fly solo, but for this test, they hitched a ride into low Earth orbit on satellites named Max Valier and Venta-1. Each satellite has one Sprite permanently riding sidecar, and the Max Valier craft has another four it could fling out into space. Unfortunately, as of August 10, ground controllers haven’t yet been able to reach the Max Valier satellite to send a “Release the Sprites!” command. One of the permanently attached Sprites — probably the one on Venta-1 — is in radio contact.

Before sending next-gen Sprites off to Alpha Centauri, scientists plan to equip them with cameras, actuators for steering and other tools. “This was really just the first step in a long journey for Starshot,” Loeb says.

With its bright hue, this snake was bound to stand out sooner or later.

A newly discovered subspecies of sea snake, Hydrophis platurus xanthos, has a narrow geographic range and an unusual hunting trick. The canary-yellow reptile hunts at night in Golfo Dulce off Costa Rica’s Pacific coast. With its body coiled up at the sea surface, the snake points its head under the water, mouth open. That folded posture “creates a buoy” that stabilizes the snake so it can nab prey in choppy water, says study coauthor Brooke Bessesen, a conservation biologist at Osa Conservation, a biodiversity-focused nonprofit in Washington, D.C. In contrast, typical Hydrophis platurus, with a black back and yellow underbelly, hunts during the day, floating straight on calm seas.
The newly described venomous snake has been reported only in a small, 320-square-kilometer area of Golfo Dulce. After analyzing 154 living and preserved specimens, the researchers described the reptile’s characteristics July 24 in Zookeys. The scientists hope that the subspecies designation will enable the Costa Rican government to protect the sunny serpent, which they worry is already at risk from overzealous animal collectors.

Imagine you’re a red-footed booby napping on a not-quite-high-enough branch of a tree. It’s nighttime on an island in the middle of the Indian Ocean, and you can’t see much of what’s around you. Then, out of the darkness comes a monster. Its claw grabs you, breaking bones and dragging you to the ground. You don’t realize it yet, but you’re doomed. The creature breaks more of your bones. You struggle, but it’s a fruitless effort. Soon the other monsters smell your blood and converge on your body, ripping it apart over the next few hours.

The monster in this horror-film scenario is a coconut crab, the world’s largest terrestrial invertebrate, which has a leg span wider than a meter and can weigh more than four kilograms.

But this is no page from a screenplay. Biologist Mark Laidre of Dartmouth University actually witnessed this scene in March 2016, during a two-month field expedition to study the crabs in the Chagos Archipelago.

Laidre, an expert on hermit crabs, had been “dying to study” their humongous cousins. Little is known about the crabs, he notes. A study earlier this year looked at the force a coconut crab’s claw can exert in the lab. But, he says, “there’s still not a single paper on how they open a coconut.”
He trekked to the remote spot in the Indian Ocean because he wanted to study the crabs in a place where few people would interfere with their natural behaviors. Laidre had heard stories that coconut crabs killed rats, and he later witnessed them munching on the rodents on the islands. “Clearly it’s in their repertoire to eat something big,” he says. And when he took inventory of the crabs’ burrows, he found the carcass of an almost full-grown red-footed booby in one. “At the time, I had assumed it was something that had died … and the crab had dragged in there,” he recalls.

But then, in the middle of the night, he saw a crab attack a bird sleeping in a tree, and he managed to catch part of the event on film. “I didn’t have the heart to videotape five coconut crabs tearing apart the bird later,” he says. “It was a little bit overwhelming. I had trouble sleeping that night.”
After the event, Laidre heard a story from a local plantation worker who had witnessed something similar a couple of years earlier. “He was sitting and eating a sandwich, and this coconut crab came right out its burrow in the middle of the daytime when … a red-footed booby… landed outside of its burrow,” Laidre says. The crab grabbed the bird’s leg and pulled it into the burrow. “The bird never emerged.”

It’s difficult to tell how often attacks like this happen, whether they’re rare or common. “Predation itself is something you don’t often witness,” Laidre says. He’d like to someday install camera traps on the islands to get a better sense of the crabs’ behavior.

But while he was in the Chagos, he did find himself in a sort of natural experiment that gave him some insight into the effect of the crabs on local bird populations. Coconut crabs live on only some of the islands. Birds can live on any of them, but their populations vary from island to island. So Laidre surveyed the islands, walking transects and counting crabs and bird nests.
“The pattern I found across the island was pronounced,” Laidre writes November 1 in Frontiers in Ecology and the Environment. On Diego Garcia, for example, a 15-kilometer transect revealed 1,000 crabs and no nesting birds. Crab-free West Island, in contrast, had an abundance of ground nests of nesting noddies.

Laidre suspects that the coconut crabs act as a “ruler of the atoll,” keeping ground-nesting bird species from finding homes on crab-filled islands. On other islands with large populations of birds, those birds might help to keep their islands crab-free by eating juvenile coconut crabs, preventing them from colonizing there.

“It’s easy to sympathize with the prey,” Laidre says, “but at the same time, there’s a lot of ecological roles that that sort of action has.”

Artificial intelligence is getting some better perspective. Like a person who can read someone else’s penmanship without studying lots of handwriting samples, next-gen image recognition AI can more easily identify familiar sights in new situations.

Made from a new type of virtual building block called capsules, these programs may cut down the enormous amount of data needed to train current image-identifying AI. And that could boost such technology as machine-made medical diagnoses, where example images may be scarce, or the responsiveness of self-driving cars, where the view is constantly shifting. Researchers with Google will present this new version of an artificial neural network at the Neural Information Processing Systems conference in Long Beach, Calif., on December 5.
Neural networks are webs of individual virtual nerve cells, or neurons, that learn to pick out objects in pictures by studying labeled example images. These networks largely classify pictures based on whether they contain certain features. For instance, a program trained on a series of head shots might conclude that a face has two eyes, a nose and a mouth. Show that program a face in profile with only one eye visible, though, and it may not recognize the photo as a face, explains Roland Memisevic, a computer scientist at the University of Montreal who was not involved in the work.

To overcome that limitation, researchers can train a neural network on millions of photos from myriad angles, and the program memorizes all the different ways a face might look. Compared with the human brain, which doesn’t need anywhere near a million examples to know what a face looks like, this system is wildly inefficient. “It’s a disaster,” Memisevic says. “Capsules try to fix that.”

Instead of webs of individual artificial neurons, these new programs have webs of clusters of neurons, called capsules. These teams of neurons can provide more information than one neuron by itself. Each capsule is designed to track not only whether a certain feature is in an image, but also properties of that feature — say, a nose’s size, orientation and position. This spatial awareness helps the program better recognize objects in previously unseen scenarios.

A capsule-containing network trained on head shots could see a face in profile and deduce — based on the appearance of the visible eye, nose and mouth — that the other eye is simply obscured, and the picture depicts a face. Since capsule networks are better at applying what they know to new situations, these neural networks need less training data to achieve the same performance as their predecessors, says Sara Sabour, a computer scientist with Google Brain in Toronto.
Sabour and her colleagues trained one capsule network on images of handwritten numbers and tested it on pictures where each number was slightly distorted. The capsule network recognized the warped images with 79 percent accuracy; a typical neural network trained on the same amount of data only got 66 percent right.

In another experiment, Sabour and colleagues trained a similar capsule network on tens of thousands of photos of toys, and then asked it to recognize the toys from new viewpoints. In this challenge, reported in a paper submitted to the 2018 International Conference on Learning Representations in Vancouver, the capsule network was wrong only about 1.4 percent of the time. A conventional neural network made almost twice as many errors.

PHILADELPHIA — Flat brains growing on microscope slides may have revealed a new wrinkle in the story of how the brain folds.

Cells inside the brains contract, while cells on the outside grow and push outward, researchers at the Weizmann Institute of Science in Rehovot, Israel, discovered from working with the lab-grown brains, or organoids. This push and pull results in folds in the organoids similar to those found in full-size brains. Orly Reiner reported the results December 5 at the joint meeting of the American Society for Cell Biology and the European Molecular Biology Organization.
Reiner and her colleagues sandwiched human brain stem cells between a glass microscope slide and a porous membrane. The apparatus allowed the cells access to nutrients and oxygen while giving the researchers a peek at how the organoids grew. The cells formed layered sheets that closed up at the edges, making the organoids resemble pita bread, Reiner said. Wrinkles began to form in the outer layers of the organoids about six days after the mini brains started growing.

These brain organoids may help explain why people with lissencephaly — a rare brain malformation in which the ridges and folds are missing — have smooth brains. The researchers used the CRISPR/Cas9 gene-editing system to make a mutation in the LIS1 gene. People with lissencephaly often have mutations in that gene. Cells carrying the mutation didn’t contract or move normally, the team found.

Reiner and her colleagues aren’t the first to propose the push-pull idea for how brains fold. But the researchers were able to show the concept at work in their experimental system, says biophysicist Xavier Trepat of the Institute for Bioengineering of Catalonia in Barcelona, who was not involved in the study. “They really were able to reproduce the shape of what we all imagine the brain should look like,” he says. “It’s not a brain, but they see structures that look like it.”

There’s both inspiring and troubling news for holiday worshippers.

Unlike other historically Christian Western nations, the United States is not losing its religion, say sociologists Landon Schnabel of Indiana University Bloomington and Sean Bock of Harvard University. But America is becoming as polarized religiously as it is politically, the researchers report online November 27 in Sociological Science.

Intense forms of religion, such as Christian evangelicalism, have maintained their popularity for nearly 30 years, Schnabel and Bock find after analyzing almost 30 years of U.S. survey data. At the same time, moderate forms of religion, such as mainline Protestantism, have consistently lost followers.
Religious moderates’ exodus from their churches stems partly from a growing link between religion and conservative politics, exemplified by the rise of the religious right in the late 1980s, the researchers suspect. Political liberals and moderates who already felt lukewarm toward the religion of their parents increasingly report identifying with no organized religion, especially if leaders of their childhood churches have taken conservative stances on social issues. Many Americans still report that they believe in God and pray, so they haven’t turned to atheism, the scientists say.

Population trends also favor intense forms of religion, Schnabel holds. Mainline Protestantism’s decline from 35 percent of the U.S. population in 1972 — about 73.5 million people — to 12 percent in 2016 — nearly 39 million people — reflects low fertility rates among these Protestants and limited numbers of new adherents from immigration and conversion. Opposite trends among U.S. evangelicals helped their form of intense Christianity surge from 18 percent of the population in 1972 to a steady level of about 28 percent from 1989 to 2016.

“More moderate forms of organized religion could become increasingly irrelevant in the United States,” Schnabel says.
The new findings play into an academic debate about the fate of religion in modern societies. Some scholars argue that in wealthy nations marked by scientific advances, religion inevitably withers. National surveys in 13 other Western, historically Christian nations show a general weakening of religious beliefs, even among intense believers, since 1991, the researchers find. But Schnabel and Bock are among those who view the United States as an exception where intense religion holds steady and even many of those leaving churches keep their faith.

The researchers examined data from nationally representative surveys on religion and other topics conducted from 1989 to 2016 by the General Social Survey, or GSS, a project of the National Opinion Research Center at the University of Chicago. GSS surveys include approximately 1,500 people annually.

Story continues below image
The proportion of the U.S. population citing strong ties to any religion held steady at around 36 percent during the study period. But the share of adults identifying themselves as religiously unaffiliated rose from around 9 percent to around 20 percent of the population, the researchers report. In another sign of loosening religious ties, those who never attended religious services rose from around 14 percent to around 25 percent of the population. Occasional attendance dropped from about 80 percent to about 70 percent.

Still, those who rarely or never prayed remained at about 24 percent of the population from 1989 to 2016. People who prayed several times a day rose from around 24 percent to about 30 percent of the total.

A belief in the Bible as God’s literal word held steady at roughly one-third of Americans. A view of the Bible as inspired by a higher power but not literal fell slightly to just under half of the population. Those tagging the Bible as a book of fables rose from around 15 percent to around 22 percent.

The new findings underscore the growing polarization of U.S. religion, say Michael Hout of New York University and Claude Fischer of the University of California, Berkeley. In a 2014 report based on GSS data, the two sociologists found that most political liberals and some political moderates who weakly identified with their parents’ religion have increasingly said that they prefer no particular religion. That trend was most pronounced for those reporting that the church they grew up with had become an advocate of politically conservative positions. Many of those people expressed a qualified belief in God, endorsing neither atheism nor absolute certainty in a higher power’s existence. Political conservatives, including those who seldom attended services or had doubts about church doctrine, had no complaints about religious leaders’ conservative political pronouncements.

Members of the millennial generation born since 1990 report low levels of religious involvement regardless of their politics, Hout adds. Millennials are skeptical of institutions in general although most still believe in God, he says. “Millennials are more comfortable with do-it-yourself religion than none at all.”

Sociologists David Voas of University College London and Mark Chaves of Duke University disagree. Millennials are part of a larger U.S. trend in which each successive generation over nearly the last century has reported slightly less intensity of religious belief than the one before, Voas and Chaves reported in a 2016 analysis of GSS data. For instance, in 2014, only 45 percent of U.S. adults ages 18 to 30 had no doubts that God exists versus 68 percent of those age 65 or over.

“The proportion of intensely religious Americans is being eroded, albeit very slowly,” Voas contends.