Researchers were peripherally aware of the fish’s existence, but “discovering” a new species requires scientific description based on specimens collected in their natural environment. Redtail garras appear to be restricted to a small stretch of river straddling the border between Thailand and Myanmar. The locality is isolated and difficult to reach, so wild redtail garras have existed in relative obscurity, despite their global appeal.

Larry Page, curator of ichthyology at the Florida Museum of Natural History, has surveyed fishes in Thailand every year since 2007. He recently encountered a few redtail garras while conducting fieldwork along the Kasat River, a tributary of the Ataran River in neighboring Myanmar.

“When we first collected specimens, we thought it must be widespread in Myanmar because of its popularity in the aquarium trade. But it turns out it’s not. It’s only in the Ataran River basin,” he said.

Page and his colleagues published the new species description in the journal Zootaxa. Redtails take their place among nearly 200 additional species in the genus Garra, one of the most diverse and widely distributed fish groups anywhere on Earth. Garra can be found in streams and rivers from the farthest reaches of western Africa through the Middle East, India and southern and eastern Asia, including parts of China. Yet their ubiquity has not resulted in a corresponding amount of research on the group.

“There’s surprisingly little information on their natural history,” Page said. There are few studies that broadly assess diversity in the genus, and little is known about relationships or how various groups within Garra have diversified.

According to Page, basic information about the fishes’ biology is available, but often in the form of studies on individual species or regions. “Most of them live in fast-moving water, and they have a disclike structure formed from a modification of the lower lip, which they use as an adhesive pad to cling to rocks and maintain their position in the water column as they feed,” he said.

Garra primarily subsist on algae and the occasional arthropod, which they eat by scraping detritus off rocks with specialized mouthparts. Like other species in the genus, redtails lack stomachs and have a snout encrusted with modified and hardened scales called tubercles. Similar structures in other fish groups are temporary; they’re used to defend nests but fall off after the breeding season has ended. Redtail tubercles are permanently attached and seem to be used as weapons, based on aggressive behavior observed in aquaria.

Redtail garra can be distinguished by their unique, elongated snout, covered in even more tubercles, which they can raise or lower, ostensibly as a means of intimidating opponents during combative encounters. And, as their common name suggests, their tails are emblazoned in red like a signal fire.

Their unique color pattern and their industrious ability to clear away algae have made them a coveted component of aquarium tanks, but their late addition to the roster of scientifically described species underscores a pressing need to take stock of biodiversity in understudied regions.

Page, who is working on a book describing the fishes of the Mae Klong River basin in Thailand, said the belated description of redtail garra is part of a larger recurring pattern. “Many of the fishes in Southeast Asia are referred to by names given to species discovered in India or Indonesia because they look similar.” People know particular species exist, Page said, but they are mistaken for those from other geographic areas, and their diversity has consequently been drastically underestimated.

Page and his colleagues named the new species Garra panitvongi after the author of another book on Thailand’s fishes, Nonn Panitvong. Lauded as a biodiversity hero by the Association of Southeast Asian Nations, Panitvong is both a businessman and an avid naturalist. He’s nurtured a passion for nature since his childhood days spent translating books about animals written in English with the help of his mom. He developed a special appreciation for Thailand’s aquatic life from his dad, who took him on fishing trips.

“Over the years, I found that in many of the places we went fishing, the environments were getting worse. There were fewer fish overall, and the fish assemblage changed for the worse, with a decreasing number of native species and more invasives,” he said.

Panitvong obtained a Master of Business Administration at the University of North Carolina Wilmington so he could effectively run his family’s sugarcane farm, but when it came time for a doctorate, he opted to study environmental science. After returning to Thailand, he created the website Siamensis.org, an open platform on which nature enthusiasts could share information about Thailand’s flora and fauna and plan excursions to wilderness areas.

In 2006, Panitvong and a friend discovered one of Thailand’s redtail garra populations and helped introduce the fish to the aquarium trade. Now, more than 15 years later, he said the official naming of redtail garra — and the discovery of species like it — help bring the world into clearer focus.

“I like to imagine humanity in a big room, painted in white. As we turned around, we would have no conception of which direction we were facing,” he said. “With each new discovery, a dot is added, a point made, and we know more and more about where we stand as a species.”

Researchers from MIT and the MIT-Watson AI Lab have developed a new, unified framework that can simultaneously predict molecular properties and generate new molecules much more efficiently than these popular deep-learning approaches.

To teach a machine-learning model to predict a molecule’s biological or mechanical properties, researchers must show it millions of labeled molecular structures — a process known as training. Due to the expense of discovering molecules and the challenges of hand-labeling millions of structures, large training datasets are often hard to come by, which limits the effectiveness of machine-learning approaches.

By contrast, the system created by the MIT researchers can effectively predict molecular properties using only a small amount of data. Their system has an underlying understanding of the rules that dictate how building blocks combine to produce valid molecules. These rules capture the similarities between molecular structures, which helps the system generate new molecules and predict their properties in a data-efficient manner.

This method outperformed other machine-learning approaches on both small and large datasets, and was able to accurately predict molecular properties and generate viable molecules when given a dataset with fewer than 100 samples.

“Our goal with this project is to use some data-driven methods to speed up the discovery of new molecules, so you can train a model to do the prediction without all of these cost-heavy experiments,” says lead author Minghao Guo, a computer science and electrical engineering (EECS) graduate student.

Guo’s co-authors include MIT-IBM Watson AI Lab research staff members Veronika Thost, Payel Das, and Jie Chen; recent MIT graduates Samuel Song ’23 and Adithya Balachandran ’23; and senior author Wojciech Matusik, a professor of electrical engineering and computer science and a member of the MIT-IBM Watson AI Lab, who leads the Computational Design and Fabrication Group within the MIT Computer Science and Artificial Intelligence Laboratory (CSAIL). The research will be presented at the International Conference for Machine Learning.

Learning the language of molecules

To achieve the best results with machine-learning models, scientists need training datasets with millions of molecules that have similar properties to those they hope to discover. In reality, these domain-specific datasets are usually very small. So, researchers use models that have been pretrained on large datasets of general molecules, which they apply to a much smaller, targeted dataset. However, because these models haven’t acquired much domain-specific knowledge, they tend to perform poorly.

The MIT team took a different approach. They created a machine-learning system that automatically learns the “language” of molecules — what is known as a molecular grammar — using only a small, domain-specific dataset. It uses this grammar to construct viable molecules and predict their properties.

In language theory, one generates words, sentences, or paragraphs based on a set of grammar rules. You can think of a molecular grammar the same way. It is a set of production rules that dictate how to generate molecules or polymers by combining atoms and substructures.

Just like a language grammar, which can generate a plethora of sentences using the same rules, one molecular grammar can represent a vast number of molecules. Molecules with similar structures use the same grammar production rules, and the system learns to understand these similarities.

Since structurally similar molecules often have similar properties, the system uses its underlying knowledge of molecular similarity to predict properties of new molecules more efficiently.

“Once we have this grammar as a representation for all the different molecules, we can use it to boost the process of property prediction,” Guo says.

The system learns the production rules for a molecular grammar using reinforcement learning — a trial-and-error process where the model is rewarded for behavior that gets it closer to achieving a goal.

But because there could be billions of ways to combine atoms and substructures, the process to learn grammar production rules would be too computationally expensive for anything but the tiniest dataset.

The researchers decoupled the molecular grammar into two parts. The first part, called a metagrammar, is a general, widely applicable grammar they design manually and give the system at the outset. Then it only needs to learn a much smaller, molecule-specific grammar from the domain dataset. This hierarchical approach speeds up the learning process.

Big results, small datasets

In experiments, the researchers’ new system simultaneously generated viable molecules and polymers, and predicted their properties more accurately than several popular machine-learning approaches, even when the domain-specific datasets had only a few hundred samples. Some other methods also required a costly pretraining step that the new system avoids.

The technique was especially effective at predicting physical properties of polymers, such as the glass transition temperature, which is the temperature required for a material to transition from solid to liquid. Obtaining this information manually is often extremely costly because the experiments require extremely high temperatures and pressures.

To push their approach further, the researchers cut one training set down by more than half — to just 94 samples. Their model still achieved results that were on par with methods trained using the entire dataset.

“This grammar-based representation is very powerful. And because the grammar itself is a very general representation, it can be deployed to different kinds of graph-form data. We are trying to identify other applications beyond chemistry or material science,” Guo says.

In the future, they also want to extend their current molecular grammar to include the 3D geometry of molecules and polymers, which is key to understanding the interactions between polymer chains. They are also developing an interface that would show a user the learned grammar production rules and solicit feedback to correct rules that may be wrong, boosting the accuracy of the system.

This work is funded, in part, by the MIT-IBM Watson AI Lab and its member company, Evonik. Paper: “Hierarchical Grammar-Induced Geometry for Data-Efficient Molecular Property Prediction”

CWD is caused by a pathogenic agent called a prion, which can pass from deer-to-deer through contact with things like prion-contaminated soil and infected bodily fluids such as urine, saliva, blood and feces. Prions, which cause disease in animals and in humans, prompt certain proteins to fold abnormally, particularly in the brain, and prevent these proteins from carrying out their normal functions. Over time, the CWD prion can cause severe brain damage and eventually death in deer.

A lot of CWD studies focus on the role soil plays in spreading the fatal neurological disease among deer. But Heather Inzalaco, a researcher in the Wisconsin Cooperative Wildlife Research Unit, housed in the UW-Madison Department of Forest and Wildlife Ecology, was curious about other potential environmental and behavioral means of transmission.

“Deer live these secret lives; we don’t see everything that they do,” Inzalaco says.

She started to consider what sort of things pester deer that could be connected to CWD: Ticks were the perfect parasitic potential culprit to investigate.

Ticks have a goal, of course, to feed on their host’s blood. Inzalaco began to wonder if ticks that acquire blood from CWD-infected deer could also host the prions and if so, could they contain enough prions to spread the disease.

The question became even more intriguing when she discovered that the most common non-aggressive social behavior that deer engage in is allogrooming.

“Deer will groom one another to get places that they can’t reach on their own through self-grooming,” Inzalaco says. “If they’re grooming each other and they’re doing that to remove ectoparasites [such as ticks], that might be problematic because they’re probably eating the ectoparasites.”

First, she needed to show that ticks can take up and harbor these prions when they feed on CWD-infected blood. She designed an experiment to do just that.

“You’d think that it’d be easy to get ticks to take a blood meal, but they are surprisingly fussy in the lab,” Inzalaco says.

She was able to determine that ticks can not only carry the prions in their blood meal, they can also carry enough of the agent to potentially infect another animal with CWD. After seeing that the phenomenon was possible in the lab, it was time to see what was happening in the wild.

Inzalaco partnered with the Department of Natural Resources to study ticks collected from deer that hunters harvested and submitted for CWD testing. Of the 176 deer with ticks she studied, 15 of the deer were also positive for CWD. Inzalaco took the ticks from the infected deer and tested the blood they contained to quantify the amount of prion the ticks harbored.

She determined that these engorged, wild ticks did carry transmissible levels of prions — just like those in the lab — making them potential mechanical vectors for the disease.

“They’re just like a little CWD tic-tac that are possibly being eaten by the deer,” Inzalaco says.

The study did not test whether prion-carrying ticks did cause transmission to other deer.

Understanding more about how CWD can spread can help improve the management of the disease. While it isn’t practical to treat all wild deer with tick preventatives, Inzalaco believes better land stewardship could help manage tick populations.

For instance, having contiguous habitat of native plant communities and properly managing areas to continue a natural fire regime has been shown to limit tick populations, she says, while more fragmented, unbalanced ecosystems riddled with invasive plants may allow ticks to proliferate more readily.

Inzalaco says it might be possible to use ticks as a way to screen for CWD in both wild and farmed deer. Current methods of diagnosis or screening involve invasive sample collection from animals or tissue sampling after their death. While testing the ticks from deer may not lead to the same level of accuracy as testing tissue samples, it could still be a useful tool to better understanding where the disease is affecting deer population in the state.

Inzalaco also believes her research can help improve the ecosystems that everyone relies on, especially the state’s hunters.

“We are all inextricably linked to ecosystem function and the biodiversity of those ecosystems,” she says. “That is really what drives my desire to learn and do good science on a daily basis. We need to make an effort to preserve our natural heritage so that we can continue living on this planet and not be overtaken by disease and have healthy animals and healthy functioning ecosystems.”

Christofer Skurka, Jessica Myrick and graduate student Yin Yang found that seeing bad news about climate change can make people more afraid over time, but it also may encourage audiences to think about what society can do to address the problem. They published the results of two separate studies in an article titled “Fanning the flames or burning out? Testing competing hypotheses about repeated exposure to threatening climate change messages,” which appeared in the journal Climatic Change.

“The public is surrounded by media coverage about climate change, and this messaging tends to be negative in tone, focusing on the threats that climate change poses to human prosperity and ecological health,” said Skurka, the paper’s lead author and an assistant professor of media studies in the Donald P. Bellisario College of Communications. “We know from years of research in the field of communication that media messaging can impact our emotions, our beliefs and, in turn, sometimes our behavior.”

The first study involved exposing participants to three days of negative news stories about climate change. A follow-up study consisted of participants reading negative news headlines about climate change in the form of Twitter posts for seven consecutive days.

“We found that three days in a row of reading doom-and-gloom news stories about climate change was linked to greater fear and less hope, which can potentially hurt an audience’s attitude that they can do anything to tackle the problem,” said Myrick, the Donald P. Bellisario Professor of Health Communication and co-funded faculty member of the Institutes of Energy and the Environment. “However, our follow-up study had people look only at headlines and not full news stories for a longer period of seven days in a row. In that study, we found that fear peaked after a few days and then held steady.”

The researchers reported that over time, people who repeatedly saw climate change headlines started to feel like they could do more to affect change and that the topic of climate change was important.

“You would think that as people are repeatedly exposed to threatening climate news devoid of solutions content that their efficacy beliefs will decrease over time,” Skurka said. “We saw the opposite pattern in our second study. People’s efficacy beliefs increased over time. In other words, the more exposure people had to these threatening news stories each day, they were increasingly likely to think that they can make a difference in addressing climate change.”

Skurka said one possibility is that as the public copes with unpleasant feelings about the enormous threat climate change presents, they may convince themselves that they have control over the situation, which translates into greater efficacy beliefs that their actions will make a difference.

“Our findings suggest that people have gotten used to doom-and-gloom reporting around climate change and what may be more important for motivating them to take action is that they see coverage of it on a daily basis,” Myrick said. “This is called an agenda-setting effect, where a topic that is covered more often in the news is then viewed as more important by people who consume the news.”

According to Skurka, decades of research in communication and psychology show that under certain circumstances, fear can be motivating.

“We found that people exposed to the high-threat headlines, which tended to evoke more fear, generally expressed greater intentions to share the information than people exposed to the low-threat headlines, which means there may be an advantage to evoking fear,” Skurka said. “However, people’s responses over time were essentially the same regardless of whether they were shown the high-threat or low-threat news headlines. That tells us that when it comes to over-time responses to repeated media exposure, simply mentioning climate change in the news activates pre-existing emotions and thoughts associated with climate change.”

Myrick added that this does not mean that fear-appeals should be used for all climate change communication. Instead, the more important factor may be communicating hope and solutions.

“For communication to be most impactful, people need to feel like there is still something we can do about it to make a difference,” Myrick said. “That should hopefully motivate reporters and strategic communicators to include information about solutions to climate change in their messaging.”