Culture

(Jena, Germany) Everything that lives has metabolites, produces metabolites and consumes metabolites. These molecules arise as intermediate and end products from chemical processes in an organism's metabolism. Therefore, they not only have huge significance for our lives, but they also provide valuable information about the condition of a living being or an environment. For example, metabolites can be used to detect diseases or, in the field of environmental technology, to examine drinking water samples. However, the diversity of these chemical compounds causes difficulties in scientific research. To date, only few molecules and their properties are known. If a sample is analysed in the laboratory, only a relatively small proportion of it can be identified, while the majority of molecules remain unknown.

Bioinformaticians at Friedrich Schiller University Jena, Germany together with colleagues from Finland and the USA, have now developed a unique method with which all metabolites in a sample can be taken into account, thus considerably increasing the knowledge gained from examining such molecules. The team reports on its successful research in the renowned scientific journal Nature Biotechnology.

Learning, recognising and assigning structural properties

"Mass spectrometry, one of the most widely used experimental methods for analysing metabolites, identifies only those molecules that can be uniquely assigned by matching them against a database. All other, previously unknown, molecules contained in the sample do not provide much information," explains Prof. Sebastian Böcker from the University of Jena. "With our newly developed method, called CANOPUS, however, we also obtain valuable insight from the unidentified metabolites in a sample, as we can assign them to existing compound classes."

CANOPUS works in two phases: first, the method generates a 'molecular fingerprint' from the fragmentation spectrum measured by means of mass spectrometry. This contains information about the structural properties of the measured molecule. In the second phase, the method uses the molecular fingerprint to assign the metabolite to a specific compound class without having to identify it.

Learning from the data

"Machine learning methods usually require large amounts of data in order to be trained. In contrast, our two-stage process makes it possible in the first step to train on a comparatively small amount of data of tens of thousands of fragmentation mass spectra. Then, in the second step, the characteristic structural properties that are significant for a compound class can be determined from millions of structures," explains Dr Kai Dührkop from the University of Jena.

The system therefore identifies these structural properties in an unknown molecule within a sample and then assigns it to a specific compound class. "This information alone is sufficient to answer many important questions," Böcker emphasises. "The precise identification of a metabolite would be far more complex and is often not necessary at all." The CANOPUS method uses a deep neural network predicting around 2,500 compound classes.

With their method, the Jena bioinformaticians have compared, for example, the intestinal flora of mice in which one experimental group had been treated with antibiotics. The examinations show which metabolites the mouse and its intestinal flora produce. Such research results can provide important information about the human digestive and metabolic system. Through two further application examples, which they present in their study, the Jena scientists demonstrate the functionality and power of the CANOPUS method.

Jena molecule search engine used millions of times

With the new method, the bioinformaticians from Jena are expanding the possibilities of the search engine for molecular structures "CSI:FingerID", which they have been making available to the international research community for around five years. Researchers around the world now use this service thousands of times a day to compare a mass spectrum from a sample with various online databases, in order to identify a metabolite more precisely. "We are approaching the one hundred millionth request and we are sure that CANOPUS will further increase the number of users," says Sebastian Böcker.

The new process strengthens the field of metabolomics - that is, research on these omnipresent small molecules - and increases its potential in many research areas, such as pharmaceuticals. Many active pharmaceutical substances in use for decades, such as penicillin, are metabolites; others could be developed with their help.

Ordinary solid-state lasers, as used in laser pointers, generate light in the visible range. For many applications, however, such as the detection of molecules, radiation in the mid-infrared range is needed. Such infrared lasers are much more difficult to manufacture, especially if the laser radiation is required in the form of extremely short, intense pulses.

For a long time, scientists have been looking for simple methods to produce such infrared laser pulses - at the TU Wien this has now been achieved, in cooperation with Harvard University. The new technology does not require large experimental setups; it can be easily miniaturized and is therefore particularly interesting for practical applications. The new results have now been presented in the journal "Nature Communications".

The frequency comb

"We generate laser light in the mid-infrared range with tailor-made quantum cascade lasers manufactured in the ultra-modern Nano-Center of TU Wien", says Johannes Hillbrand from the Institute of Solid State Electronics at the TU Vienna, first author of the study. While in ordinary solid-state lasers the type of light emitted depends on the atoms in the material, in quantum cascade lasers tiny structures in the nanometer range are crucial. By designing these structures appropriately, the wavelength of the light can be precisely adjusted.

"Our quantum cascade lasers do not just deliver a single color of light, but a whole range of different frequencies," says Ass.Prof. Benedikt Schwarz, who led the research work in his ERC grant funded project. "These frequencies are arranged very regularly, always with the same distance in between, like the teeth of a comb. Therefore, such a spectrum is called a frequency comb".

Light is like a pendulum

However, it is not only the frequencies emitted by such a quantum cascade laser that are decisive, but also the phase with which the respective light waves oscillate. "You can compare this to two pendulums connected by a rubber band," explains Johannes Hillbrand. "They can either swing back and forth, exactly in parallel, or opposite to each other, so that they either swing towards each other or away from each other. And these two vibration modes have slightly different frequencies."

It is quite similar with laser light, which is composed of different wavelengths: The individual light waves of the frequency comb can oscillate exactly in sync - then they superimpose each other in an optimal way and can generate short, intense laser pulses. Or there can be shift between their oscillations, in which case no pulses are created, but laser light with an almost continuous intensity.

The light modulator

"In quantum cascade lasers, it has previously been difficult to switch back and forth between these two variants," says Johannes Hillbrand. "However, we have built a tiny modulator into our quantum cascade laser, which the light waves pass by again and again." An alternating electrical voltage is applied to this modulator. Depending on the frequency and strength of the voltage, different light oscillations can be excited in the laser.

"If you drive this modulator at exactly the right frequency, you can achieve that the different frequencies of our frequency comb all oscillate at exactly in sync," says Benedikt Schwarz. "This makes it possible to combine these frequencies into short, intense laser pulses - more than 12 billion times per second".

This level of control over short infrared laser pulses was previously not possible with semiconductor lasers. Similar types of light could at best only be generated using very expensive and lossy methods. "Our technology has the decisive advantage that it can be miniaturized," emphasizes Benedikt Schwarz. "One could use it to build compact measuring instruments that use these special laser beams to search for very specific molecules in a gas sample, for example. Thanks to the high light intensity of the laser pulses, measurements that require two photons at the same time are also possible.

Resistive switching memory devices offer several advantages over the currently used computer memory technology. Researchers from the MIPT Atomic Layer Deposition Lab have joined forces with colleagues from Korea to study the impact of electrode surface morphology on the properties of a resistive switching memory cell. It turned out that thicker electrodes have greater surface roughness and are associated with markedly better memory cell characteristics. The research findings were published in ACS Applied Materials & Interfaces.

Some materials, such as transition metal oxides, can switch from a dielectric to a conductive state and back under applied voltage. This effect underlies resistive random-access memory, a highly promising technology for nonvolatile storage. RRAM devices based on transition metal oxides are characterized by low energy consumption, great endurance, ease of extension, and rapid operation, prompting many companies to invest in the technology.

A resistive memory cell is a layered structure with an insulating layer positioned between two electrodes, to which the switching voltage is applied. The properties of the cell depend on the material between the electrodes, as well as on the composition and shape of the electrodes themselves. It is common for one electrode to be made of titanium nitride and the other of platinum. However, platinum is incompatible with modern semiconductor technology due to the absence of dry etching capability. This is not the case with ruthenium, which has a further advantage of being suitable for atomic layer deposition (ALD), enabling the manufacture of 3D vertical memory structures.

Study co-author and MIPT PhD student Aleksandra Koroleva from the University's School of Electronics, Photonics and Molecular Physics commented: "To investigate how electrode thickness affects memory cell parameters, we grew ruthenium electrodes with a varying number of atomic layer deposition cycles. We then examined the surface of the electrodes using atomic force microscopy." The team found that as the number of ALD layers grew, the grain size on the electrode surface increased from 5 to 70 nanometers.

The researchers tested the performance of their ruthenium films with different thicknesses as the bottom electrode in tantalum oxide-based RRAM, showing that thicker -- and therefore rougher -- electrodes actually improved the key performance characteristics of the memory device: its stability and endurance. Increasing ruthenium film thickness resulted in a lower memory cell resistance in both states and a higher resistance ratio between the low- and high-resistance states. Enhancing electrode roughness also decreased the forming and switching voltages, and increased the device's endurance to an impressive 50 million switching cycles.

To explain their findings, the team proposed a simplified model that reflects the electric field distribution on large grains on the ruthenium electrode surface. The explanation was confirmed with conductive atomic force microscopy.

"Our findings offer insights into how memory cells of the new type could be greatly improved. Thicker ruthenium films used as electrodes have rougher surfaces. This in turn gives rise to areas of locally enhanced electric field on the slopes of the grains that boost the key performance characteristics of the device. We believe that our investigation will help to create more efficient and reliable memory devices in the future," adds study co-author Andrey Markeev, who leads the ALD group at MIPT.

FRANKFURT. The surface of the SARS-CoV-2 virus is studded with spike proteins. The virus needs these in order to dock onto proteins (ACE2 receptors) on the surface of the host cell. Before this docking is possible, parts of the spike protein have to be cleaved by the host cell's enzymes - proteases.

In cell culture experiments with various cell types, the international scientific team led by Professor Jindrich Cinatl, Institute for Medical Virology at the University Hospital Frankfurt, Professor Martin Michaelis, and Dr Mark Wass (both University of Kent) demonstrated that the protease inhibitor aprotinin can inhibit virus replication by preventing SARS-CoV2 entry into host cells. Moreover, aprotinin appears to compensate for a SARS-CoV2-induced reduction of endogenous protease inhibitors in virus-infected cells.

Influenza viruses require host cell proteases for cell entry in a similar way as coronaviruses. Hence, an aprotinin aerosol is already approved in Russia for the treatment of influenza.

Professor Jindrich Cinatl said: "Our findings show that aprotinin is effective against SARS-CoV2 in concentrations that can be achieved in patients. In aprotinin we have a drug candidate for the treatment of COVID-19 that is already approved for other indications and could readily be tested in patients."

Galaxies like the Milky Way formed by the merging of smaller progenitor galaxies. An international team of astrophysicists led by Dr Diederik Kruijssen from the Centre for Astronomy at Heidelberg University has succeeded in reconstructing the merger history of our home galaxy, creating a complete family tree. To achieve this, the researchers analysed the properties of globular clusters orbiting the Milky Way with artificial intelligence. Their investigations revealed a previously unknown galaxy collision that must have permanently altered the appearance of the Milky Way.

Globular clusters are dense groups of up to a million stars that are almost as old as the universe itself. The Milky Way hosts over 150 of such clusters. "Many of them came from smaller galaxies that later merged to form the Milky Way that we live in today," explains Dr Kruijssen. To study the merger history, the Heidelberg researcher and his colleague Dr Joel Pfeffer of Liverpool John Moores University (United Kingdom) and their research groups developed a suite of advanced computer simulations, called E-MOSAICS. These simulations include a complete model for the formation, evolution, and destruction of globular clusters.

The German-British team used these simulations to relate the ages, chemical compositions, and orbital motions of the globular clusters to the properties of the progenitor galaxies in which they formed, more than ten billion years ago. By applying these insights to groups of globular clusters in the Milky Way, they not only determined how massive these progenitor galaxies were, but also when they merged with our home galaxy.

"The main challenge was that the merger process is extremely messy, because the orbits of the globular clusters are completely reshuffled," explains Dr Kruijssen. "To overcome this complexity, we developed an artificial neural network and trained it on the E-MOSAICS simulations. We were astonished at how precisely the artificial intelligence allowed us to reconstruct the merger histories of the simulated galaxies, using only their globular clusters." The researchers then applied the neural network to groups of globular clusters in the Milky Way and precisely determined the stellar masses and merger times of the progenitor galaxies. They also discovered a previously unknown collision between the Milky Way and an unknown galaxy, which the researchers named "Kraken".

"The collision with Kraken must have been the most significant merger the Milky Way ever experienced," Dr Kruijssen adds. Before, it was thought that a collision with the Gaia-Enceladus galaxy some nine billion years ago was the biggest collision event. However, the merger with Kraken took place eleven billion years ago, when the Milky Way was four times less massive than today. "As a result, the collision with Kraken must have truly transformed what the Milky Way looked like at the time," explains the Heidelberg scientist.

Taken together, these findings allowed the team of researchers to reconstruct the first complete family tree of our home galaxy. Over the course of its history, the Milky Way cannibalised about five galaxies with more than 100 million stars, and about ten more with at least ten million stars. The most massive progenitor galaxies collided with the Milky Way between six and eleven billion years ago. Dr Kruijssen expects that these predictions will aid the future search for the remains of the progenitor galaxies. "The debris of more than five progenitor galaxies has now been identified. With current and upcoming telescopes, it should be possible to find them all," the Heidelberg researcher concludes.

The research results were published in Monthly Notices of the Royal Astronomical Society.

ATHENS, Ohio (Nov. 23, 2020) - Dr. Jeff Russell, associate professor of athletic training within the College of Health Sciences and Professions at Ohio University, is shining a light on a segment of concussion patients who often go unnoticed in comparison to athletes: performing artists.

Russell's new paper highlights the risk of concussion for dance, circus, theater and film and television stunt performers, along with guidelines for treatment. It is the first-time concussion risk for film and television stunt performers that is being highlighted in scientific literature.

"When you stop and think about how influential a field is, pretty much everybody in the world watches movies and television. This type of medium is an art form with a huge influence, but people don't know what goes on behind the scenes to make it look so cool on television or on the movie screen," Russell said. "That's where my role as a healthcare worker in performing arts comes into play. I understand what artists go through - I look at what they do with different eyes, and I look at it in terms of risk and what potential injuries could happen or what might be an unsafe condition."

The article was released by Physical Medicine and Rehabilitation Clinics of North America on Oct. 29, 2020. The key findings include that dance and performing arts are highly physical activities and performers could experience a head impact from many sources, yet the scientific literature devoted to concussions in performing arts is very low in comparison to that in sports.

Russell noted that this may be due to the fact that, while everyone watches sports and can see injuries occur, injuries to stunt actors and performing artists often occur behind the scenes and away from the public view.

"When football players are running into each other, when hockey players are skating into each other, when soccer players are heading the ball, it's obvious to us that concussions can happen, but that is not obvious in the arts," Russell said. "The injuries aren't necessarily happening only during the performances because that's just one small piece of artistic effort. Injuries happen a lot in the rehearsals or backstage behind the scenes where you can't see them."

Those with sport concussions also receive better care because their injuries are more well known. In fact, many athletes have athletic trainers on hand who are equipped to recognize concussion symptoms; however, performing artists, including stunt performers, do not always have the same healthcare resources.

"It's quite unfortunate," Russell said, noting the lack of adequate concussion care for performing artists. "We've got to change the way things are done, with the goal of protecting the performers while not impeding the production of the film or television program and not negatively affecting the profitability of the production companies."

Stunt performers may also be more prone to remain quiet about an injury in fear of losing their job. Russell compared it to the NFL, where athletes won't receive a pay cut if they have to sit out of games while on concussion protocol. However, for the stunt performers, if they cannot perform their stunts, they are out of a job. This means that their symptoms can go undiagnosed and untreated, which can be dangerous.

"Concussions are important primarily because they're an injury to your brain. If you're not taking care of your brain, that means the rest of your body won't work right," Russell said. "If you have a concussion and you don't tell anybody, and keep doing what you're doing then you have another concussion on top of it, that can be a very, very dangerous situation. One of the side effects of doubling concussions can even be death - the brain just can't handle it and shuts down."

Director of OHIO's School of Theater Merri Biechler said that Russell's research is changing the way that the performing arts industry views and addresses injuries.

"Many theater professionals push through injury because there's a false sense that 'the show must go on.' Dr. Russell's work requires us to slow down, to focus on self care," Biechler said.

Russell has been working with concussions in performing artists since the early 2000s after a student came to him and requested help. Since then, he has led research into the causes and risks of the injury in performers. He is also the director of OHIO's Science and Health in Artistic Performance (SHAPe) Clinic, a facility that provides injured performing artists care from licensed athletic trainers.

"This is why I changed my career from sports medicine to performing arts medicine. I saw dancers and some other artists that were suffering injuries, but there was nobody to take care of them. It didn't sit right with me," Russell said. "I can't solve all the world's problems, but I can work right here and take care of these people. So that's what we're doing now at OHIO with the SHAPe Clinic. We can't do it for everyone, but we're going to do it for these ones here at our University. We're going to set an example, and we're going to move forward and be ones that are on the cutting edge of this. Both the clinical care and the research are fundamental to this mission."

OHIO's SHAPe Clinic is a partnership between College of Fine Arts and the College of Health Sciences and Professions, and provides care to many students in the performing arts, including dance, music, theater and the Ohio University Marching 110.

"The work being done by the SHAPe Clinic to not only support the physical needs of our performance students in theater, dance and music, but also to bring attention to the importance of access to this type of expert care and oversight in the field is extraordinary," College of Fine Arts Dean Dr. Matthew Shaftel said. "This early socialization of often overlooked specialized care for performing artists makes a world of difference when these students transition to the professional stage and screen."

This interdisciplinary collaboration has already made great strides in increasing the safety of OHIO's theater performers.

"A few years back, an actor experienced a concussion during a performance. We allowed the actor to perform future shows within the concussion protocol and informed the audience that they would see a slightly different physical performance. The production experienced a heightened attention that literally generated something new on the stage," Biechler said. "It was a great lesson that when we take care of our health, unique experiences are created."

Russell hopes that his research work will continue to shine a light on the subpar care many performing artists--not just stunt performers--receive for concussions, and that eventually this will lead to changes in the industry and the care artists receive.

"Ultimately, this work is about helping people," Russell said. "It's about serving them, lifting them up, adding value to them and helping them, and if that is what your research is about, I don't think you can do any better than that."

Cambridge, MA (November 23, 2020)--An international team of scientists led by the Center for Astrophysics | Harvard & Smithsonian, has identified a problem with the growing interest in extractable resources on the moon: there aren't enough of them to go around. With no international policies or agreements to decide "who gets what from where," scientists believe tensions, overcrowding, and quick exhaustion of resources to be one possible future for moon mining projects. The paper published today in the Philosophical Transactions of the Royal Society A.

"A lot of people think of space as a place of peace and harmony between nations. The problem is there's no law to regulate who gets to use the resources, and there are a significant number of space agencies and others in the private sector that aim to land on the moon within the next five years," said Martin Elvis, astronomer at the Center for Astrophysics | Harvard & Smithsonian and the lead author on the paper. "We looked at all the maps of the Moon we could find and found that not very many places had resources of interest, and those that did were very small. That creates a lot of room for conflict over certain resources."

Resources like water and iron are important because they will enable future research to be conducted on, and launched from, the moon. "You don't want to bring resources for mission support from Earth, you'd much rather get them from the Moon. Iron is important if you want to build anything on the moon; it would be absurdly expensive to transport iron to the moon," said Elvis. "You need water to survive; you need it to grow food--you don't bring your salad with you from Earth--and to split into oxygen to breathe and hydrogen for fuel."

Interest in the moon as a location for extracting resources isn't new. An extensive body of research dating back to the Apollo program has explored the availability of resources such as helium, water, and iron, with more recent research focusing on continuous access to solar power, cold traps and frozen water deposits, and even volatiles that may exist in shaded areas on the surface of the moon. Tony Milligan, a Senior Researcher with the Cosmological Visionaries project at King's College London, and a co-author on the paper said, "Since lunar rock samples returned by the Apollo program indicated the presence of Helium-3, the moon has been one of several strategic resources which have been targeted."

Although some treaties do exist, like the 1967 Outer Space Treaty--prohibiting national appropriation--and the 2020 Artemis Accords--reaffirming the duty to coordinate and notify--neither is meant for robust protection. Much of the discussion surrounding the moon, and including current and potential policy for governing missions to the satellite, have centered on scientific versus commercial activity, and who should be allowed to tap into the resources locked away in, and on, the moon. According to Milligan, it's a very 20th century debate, and doesn't tackle the actual problem. "The biggest problem is that everyone is targeting the same sites and resources: states, private companies, everyone. But they are limited sites and resources. We don't have a second moon to move on to. This is all we have to work with." Alanna Krolikowski, assistant professor of science and technology policy at Missouri University of Science and Technology (Missouri S&T) and a co-author on the paper, added that a framework for success already exists and, paired with good old-fashioned business sense, may set policy on the right path. "While a comprehensive international legal regime to manage space resources remains a distant prospect, important conceptual foundations already exist and we can start implementing, or at least deliberating, concrete, local measures to address anticipated problems at specific sites today," said Krolikowski. "The likely first step will be convening a community of prospective users, made up of those who will be active at a given site within the next decade or so. Their first order of business should be identifying worst-case outcomes, the most pernicious forms of crowding and interference, that they seek to avoid at each site. Loss aversion tends to motivate actors."

There is still a risk that resource locations will turn out to be more scant than currently believed, and scientists want to go back and get a clearer picture of resource availability before anyone starts digging, drilling, or collecting. "We need to go back and map resource hot spots in better resolution. Right now, we only have a few miles at best. If the resources are all contained in a smaller area, the problem will only get worse," said Elvis. "If we can map the smallest spaces, that will inform policymaking, allow for info-sharing and help everyone to play nice together so we can avoid conflict."

While more research on these lunar hot spots is needed to inform policy, the framework for possible solutions to potential crowding are already in view. "Examples of analogs on Earth point to mechanisms for managing these challenges. Common-pool resources on Earth, resources over which no single actor can claim jurisdiction or ownership, offer insights to glean. Some of these are global in scale, like the high seas, while other are local like fish stocks or lakes to which several small communities share access," said Krolikowski, adding that one of the first challenges for policymakers will be to characterize the resources at stake at each individual site. "Are these resources, say, areas of real estate at the high-value Peaks of Eternal Light, where the sun shines almost continuously, or are they units of energy to be generated from solar panels installed there? At what level can they can realistically be exploited? How should the benefits from those activities be distributed? Developing agreement on those questions is a likely precondition to the successful coordination of activities at these uniquely attractive lunar sites."

Oversized, hairy tarantulas may be unsightly and venomous, but surprisingly their hunter toxin may hold answers to better control of chronic pain.

A bird-catching Chinese tarantula bite contains a stinger-like poison that plunges into a molecular target in the electrical signaling system of their prey's nerve cells.

A new high-resolution cryo-electron microscopy study shows how the stinger quickly locks the voltage sensors on sodium channels, the tiny pores on cell membranes that create electrical currents and generate signals to operate nerves and muscles. Trapped in their resting position, the voltage sensors are unable to activate.

The findings are published Nov. 23 in Molecular Cell, a journal of Cell Press.

"The action of the toxin has to be immediate because the tarantula has to immobilize its prey before it takes off," said William Catterall, professor of pharmacology at the University of Washington School of Medicine. He was the senior researcher, along with pharmacology professor and Howard Hughes Medical Institute investigator, Ning Zheng, on the study of the molecular damage inflicted by tarantula venom.

While some might dismiss those tarantulas as ugly, tough and mean, medical scientists are actually interested in their venom's ability to trap the resting state of the voltage sensor on voltage-gated sodium channels and shut them down. Such studies of toxins from these "big, nasty dudes," as Catterall describes them, could point to new approaches to structurally designing drugs that might treat chronic pain by blocking sensory nerve signals.

Catterall explained that chronic pain is a difficult-to-treat disorder. Efforts to seek relief can sometimes be a gateway to opiate overdose, addiction, prolonged withdrawal, and even death. The development of safer, more effective, non-addictive drugs for pain management is a vital need.

However, because it has been hard to capture the functional form of the tarantula toxin-ion channel chemical complex, reconstructing the toxin's blocking method in a small molecule has so far eluded molecular biologists and pharmacologists seeking new ideas for better pain drug designs.

Researchers overcame this obstacle by engineering a chimeric model sodium channel. Like mythical centaurs, chimeras are composed of parts of two or more species. The researchers took the toxin-binding region from a specific type of human sodium channel that is crucial for pain transmission and imported it into their model ancestral sodium channel from a bacterium. They were then able to obtain a clear molecular view of configuration of the potent toxin from tarantula venom as it binds tightly to its receptor site on the sodium channel.

This achievement revealed the structural basis for voltage sensor trapping of the resting state of the sodium channel by this toxin.

"Remarkably, the toxin plunges a 'stinger' lysine residue into a cluster of negative charges in the voltage sensor to lock it in place and prevent its function," Catterall said. "Related toxins from a wide range of spiders and other arthropod species use this molecular mechanism to immobilize and kill their prey."

Catterall explained the medical research importance of this discovery. The human sodium channel placed into the chimeric model is called the Nav1.7 channel. It plays an essential role, he noted, in transmission of pain information from the peripheral nervous system to the spinal cord and brain and is therefore a prime target for pain therapeutics.

"Our structure of this potent tarantula toxin trapping the voltage sensor of Nav1.7 in the resting state," Catterall noted, "provides a molecular template for future structure-based drug design of next-generation pain therapeutics that would block function of Nav1.7 sodium channels."

Columbus, Ohio - A study led by researchers at The Ohio State University Comprehensive Cancer Center - Arthur G. James Cancer Hospital and Richard J. Solove Research Institute (OSUCCC - James) described a potential therapeutic anticancer vaccine that frees suppressed cancer-killing immune cells, enabling them to attack and destroy a tumor.

Published in the journal Oncoimmunology, on October 1, 2020, the findings showed that the peptide called PD1-Vaxx, a first checkpoint inhibitor vaccine, was safe and effective in a colon cancer syngeneic animal model.

The vaccine produced polyclonal antibodies that inhibit the programmed cell death receptor, PD-1, on cancer cells. The vaccine mimics the action of the PD-1 inhibitor nivolumab (pronounced nih-VOL-yoo-mab, marketed as Opdivo), but it avoids triggering the innate and acquired resistance associated with that and related agents, the researchers say.

The study found that PD1-Vaxxwas effective in inhibiting tumor growth. It was even more effective when used in combination with a second therapeutic peptide vaccine, one that targets two sites on the HER-2 receptor on colon cancer cells. The combination treatment produced complete responses in nine of 10 animals. That vaccine, called B-Vaxx, was developed earlier by the same research team.

"Our study is important for two key reasons," says first author and vaccine developer Pravin T. P. Kaumaya, PhD, a member of the OSUCCC - James Translational Therapeutics Research Program and professor of medicine at The Ohio State College of Medicine. "First, PD1-Vaxx activates both B- and T-cell functions to promote tumor clearance. Second, the treatment is targeted to block signaling pathways that are crucial for tumor growth and maintenance. By giving this vaccine in combination with an immunotherapy drug, we are essentially super-charging and specifically directing the immune system to target and kill cancer cells."

Like the immune therapy drug nivolumab, PD1-Vaxx is an immune checkpoint inhibitor. Immune checkpoints are proteins that keep immune cells from attacking healthy body cells. PD-1 is a checkpoint protein on killer T cells. PD-L1 is another checkpoint protein that is on healthy cells and on many cancer cells. When PD-1 on T cells binds with PD-L1 on a body cell or a cancer cell, it suppresses the T cell, preventing it from killing the cell.

Nivolumab works by blocking PD-1 from binding with PD-L1, thereby allowing T cells to kill a patient's cancer cells. But while nivolumab consists of anti-PD-1 monoclonal antibodies, which target a single location on the PD-1 protein, the experimental vaccine PD1-Vaxx triggers a range of antibodies - a polyclonal antibody response - that blocks multiple sites on PD-1 and could more effectively inhibit the protein.

For this study, Kaumaya and his colleagues used cell lines and animal models to evaluate four PD-1 B-cell peptide epitopes as vaccine candidates. Of these, the PD-1 epitope sequence 92-110 significantly reduced tumor growth in an animal colon cancer tumor model and was chosen for the PD1-Vaxx inhibitory vaccine.

Key findings:

PD1-Vaxx outperformed the standard anti-mouse PD-1 antibody (mAb 29F.1A12) in an animal model of HER-2 expressing colon carcinoma;

The combination of PD1-Vaxx with combo HER-2 peptide vaccine (B-Vaxx) showed enhanced inhibition of tumor growth in a HER-2-positive colon cancer model;

Both the PD-1 and the combined vaccines were safe with no evidence of toxicity or autoimmunity.

"With additional study," Kaumaya says, "we believe PD1-Vaxx will prove to be safer, more effective and have a lower incidence of resistance than checkpoint-blockade antibodies."

This study was supported by grants from the National Institutes of Health (CA84356, CA13508, CA181115), and by Imugene Ltd. The safety of the vaccine was confirmed in pre-clinical animal studies at OSU and Charles River labs (Ashland, Ohio).

Vaccine Received IND Approval

In November 2020, the U.S. Food and Drug Administration (FDA) granted investigational new drug (IND) approval to Imugene for clinical testing of the investigational vaccine, known as PD1-Vaxx, an important milestone in the research collaboration between Ohio State and Imugene.

A first-in-human, phase1 clinical trial to test the vaccine is expected to open at the OSUCCC James in early 2021 for certain patients with non-small cell lung cancer. Additional U.S. sites may be added to trial at a later date.

"We are excited to begin testing of this vaccine in the United States to offer new hope to patients with lung and other cancers. Reaching this point where we can transition our findings from the lab to the clinic speaks to the perseverance and dedication of Imugene's clinical and research team -- including our research lab staff at Ohio State -- to build on the clinical and commercial potential," said Kaumaya.

Imugene's Chief Executive Officer & Managing Director Leslie Chong, said "the multiple commercial, strategic and clinical benefits of our collaboration with the OSU secures our leadership position in the promising B-cell peptide cancer vaccine sector, and in particular PD-1 checkpoint inhibitors, where OSU's pre-clinical work for a Phase I PD-1 clinical trial was pivotal to our FDA IND approval."

"This collaborative research with Imugene has closely paralleled my personal work over the past two decades, and together we form a strong team driving multiple combination immunotherapy drugs through the clinic targeting lung, breast, gastric and other cancer targets. This collaborative venture with Imugene supports the rapid development to achieving a potential cure for several important cancer targets."

CAMBRIDGE, MA -- Since the coronavirus pandemic began in the spring, many people have only seen their close friends and loved ones during video calls, if at all. A new study from MIT finds that the longings we feel during this kind of social isolation share a neural basis with the food cravings we feel when hungry.

The researchers found that after one day of total isolation, the sight of people having fun together activates the same brain region that lights up when someone who hasn't eaten all day sees a picture of a plate of cheesy pasta.

"People who are forced to be isolated crave social interactions similarly to the way a hungry person craves food. Our finding fits the intuitive idea that positive social interactions are a basic human need, and acute loneliness is an aversive state that motivates people to repair what is lacking, similar to hunger," says Rebecca Saxe, the John W. Jarve Professor of Brain and Cognitive Sciences at MIT, a member of MIT's McGovern Institute for Brain Research, and the senior author of the study.

The research team collected the data for this study in 2018 and 2019, long before the coronavirus pandemic and resulting lockdowns. Their new findings, described today in Nature Neuroscience, are part of a larger research program focusing on how social stress affects people's behavior and motivation.

Former MIT postdoc Livia Tomova, who is now a research associate at Cambridge University, is the lead author of the paper. Other authors include Kimberly Wang, a McGovern Institute research associate; Todd Thompson, a McGovern Institute scientist; Atsushi Takahashi, assistant director of the Martinos Imaging Center; Gillian Matthews, a research scientist at the Salk Institute for Biological Studies; and Kay Tye, a professor at the Salk Institute.

Social craving

The new study was partly inspired by a recent paper from Tye, a former member of MIT's Picower Institute for Learning and Memory. In that 2016 study, she and Matthews, then an MIT postdoc, identified a cluster of neurons in the brains of mice that represent feelings of loneliness and generate a drive for social interaction following isolation. Studies in humans have shown that being deprived of social contact can lead to emotional distress, but the neurological basis of these feelings is not well-known.

"We wanted to see if we could experimentally induce a certain kind of social stress, where we would have control over what the social stress was," Saxe says. "It's a stronger intervention of social isolation than anyone had tried before."

To create that isolation environment, the researchers enlisted healthy volunteers, who were mainly college students, and confined them to a windowless room on MIT's campus for 10 hours. They were not allowed to use their phones, but the room did have a computer that they could use to contact the researchers if necessary.

"There were a whole bunch of interventions we used to make sure that it would really feel strange and different and isolated," Saxe says. "They had to let us know when they were going to the bathroom so we could make sure it was empty. We delivered food to the door and then texted them when it was there so they could go get it. They really were not allowed to see people."

After the 10-hour isolation ended, each participant was scanned in an MRI machine. This posed additional challenges, as the researchers wanted to avoid any social contact during the scanning. Before the isolation period began, each subject was trained on how to get into the machine, so that they could do it by themselves, without any help from the researcher.

"Normally, getting somebody into an MRI machine is actually a really social process. We engage in all kinds of social interactions to make sure people understand what we're asking them, that they feel safe, that they know we're there," Saxe says. "In this case, the subjects had to do it all by themselves, while the researcher, who was gowned and masked, just stood silently by and watched."

Each of the 40 participants also underwent 10 hours of fasting, on a different day. After the 10-hour period of isolation or fasting, the participants were scanned while looking at images of food, images of people interacting, and neutral images such as flowers. The researchers focused on a part of the brain called the substantia nigra, a tiny structure located in the midbrain, which has previously been linked with hunger cravings and drug cravings. The substantia nigra is also believed to share evolutionary origins with a brain region in mice called the dorsal raphe nucleus, which is the area that Tye's lab showed was active following social isolation in their 2016 study.

The researchers hypothesized that when socially isolated subjects saw photos of people enjoying social interactions, the "craving signal" in their substantia nigra would be similar to the signal produced when they saw pictures of food after fasting. This was indeed the case. Furthermore, the amount of activation in the substantia nigra was correlated with how strongly the patients rated their feelings of craving either food or social interaction.

Degrees of loneliness

The researchers also found that people's responses to isolation varied depending on their normal levels of loneliness. People who reported feeling chronically isolated months before the study was done showed weaker cravings for social interaction after the 10-hour isolation period than people who reported a richer social life.

"For people who reported that their lives were really full of satisfying social interactions, this intervention had a bigger effect on their brains and on their self-reports," Saxe says.

The researchers also looked at activation patterns in other parts of the brain, including the striatum and the cortex, and found that hunger and isolation each activated distinct areas of those regions. That suggests that those areas are more specialized to respond to different types of longings, while the substantia nigra produces a more general signal representing a variety of cravings.

Now that the researchers have established that they can observe the effects of social isolation on brain activity, Saxe says they can now try to answer many additional questions. Those questions include how social isolation affect people's behavior, whether virtual social contacts such as video calls help to alleviate cravings for social interaction, and how isolation affects different age groups.

The researchers also hope to study whether the brain responses that they saw in this study could be used to predict how the same participants responded to being isolated during the lockdowns imposed during the early stages of the coronavirus pandemic.

CHAPEL HILL, NC - Strong emotions such as fear and anxiety tend to be accompanied and reinforced by measurable bodily changes including increased blood pressure, heart rate and respiration, and dilation of the eyes' pupils. These so-called "physiological arousal responses" are often abnormally high or low in psychiatric illnesses such as anxiety disorders and depression. Now scientists at the UNC School of Medicine have identified a population of brain cells whose activity appears to drive such arousal responses.

The scientists, whose study is published in Cell Reports, found that artificially forcing the activity of these brain cells in mice produced an arousal response in the form of dilated pupils and faster heart rate, and worsened anxiety-like behaviors.

The finding helps illuminate the neural roots of emotions, and point to the possibility that the human-brain counterpart of the newly identified population of arousal-related neurons might be a target of future treatments for anxiety disorders and other illnesses involving abnormal arousal responses.

"Focusing on arousal responses might offer a new way to intervene in psychiatric disorders," said first author Jose Rodríguez-Romaguera, PhD, assistant professor in the UNC Department of Psychiatry and member of the UNC Neuroscience Center, and co-director of the Carolina Stress Initiative at the UNC School of Medicine.

Rodríguez-Romaguera and co-first author Randall Ung, PhD, an MD-PhD student and adjunct assistant professor in the Department of Psychiatry, led this study when they were members of the UNC laboratory of Garret Stuber, PhD, who is now at the University of Washington.

"This work not only identifies a new population of neurons implicated in arousal and anxiety, but also opens the door for future experiments to systematically examine how molecularly defined cell types contribute to complex emotional and physiological states," Stuber said. "This will be critical going forward for developing new treatments for neuropsychiatric disorders."

Anxiety disorders, depression, and other disorders featuring abnormally high or low arousal responses affect a large fraction of the human population, including tens of millions of adults in the United States alone. Treatments may alleviate symptoms, but many have adverse side effects, and the root causes of these disorders generally remain obscure.

Untangling these roots amid the complexity of the brain has been an enormous challenge, one that laboratory technology has only recently begun to surmount.

Rodríguez-Romaguera, Ung, Stuber and colleagues examined a brain region within the amygdala called the BNST (bed nucleus of the stria terminalis), which has been linked in prior research to fear and anxiety-like behaviors in mice.

Increasingly, scientists view this region as a promising target for future psychiatric drugs. In this case, the researchers zeroed in on a set of BNST neurons that express a neurotransmitter gene, Pnoc, known to be linked to pain sensitivity and more recently to motivation.

The team used a relatively new technique called two-photon microscopy to directly image BNST Pnoc neurons in the brains of mice while the mice were presented with noxious or appealing odors - stimuli that reliably induce fear/anxiety and reward behaviors, respectively, along with the appropriate arousal responses. In this way, the scientists found that activity in these neurons tended to be accompanied by the rapid dilation of the pupils of the mice when the animals were presented with either of these odor stimuli.

The researchers then used another advanced technique called optogenetics - using light to control genetically engineered cells - to artificially drive the activity of the BNST Pnoc neurons. They found that spurring on BNST Pnoc activity triggered a pupillary response, as well as increased heart rate. Optogenetically driving the neurons while the mice underwent an anxiety-inducing maze test (traditionally used to assess anxiety drugs) increased the animals' signs of anxiety, while optogenetically quieting the neurons had the opposite effect.

"Essentially we found that activating these BNST Pnoc neurons drives arousal responses and worsens anxiety-like states," Rodríguez-Romaguera said.

The discovery is mainly a feat of basic neuroscience. But it also suggests that targeting arousal-driving neurons such as BNST Pnoc neurons with future drugs might be a good way to reduce abnormally strong responses to negative stimuli in anxiety disorders, for example, and to boost abnormally weak responses to positive stimuli in depression.

The study uncovered evidence that BNST Pnoc neurons are not all the same but differ in their responses to positive or negative stimuli, and the researchers are now cataloguing these BNST Pnoc neuron sub-groups.

"Even this small part of the amygdala is a complex system with different types of neurons," Ung said. Teasing this apart will help us understand better how this system works."

In the most comprehensive analysis to date of U.S. children tested and treated for COVID-19, an organization representing seven of the nation's largest pediatric medical centers reports that some groups of children are faring significantly worse than children in general during the pandemic.

Findings from the PEDSnet organization--which includes Cincinnati Children's--were published Nov. 23, 2020, in JAMA Pediatrics. The report is based on electronic medical records data from more than 135,000 children who have been tested for infections from the SARS-CoV-2 virus from Jan. 1 through Sept. 8, 2020.

"These findings are important because they improve our understanding of the impact of COVID-19 in the pediatric population," says Nathan Pajor, MD, a pulmonary medicine specialist at Cincinnati Children's and a co-author of the study. "We see that relative to adults, kids are less likely to have severe disease or to die from COVID-19. However, we also notice the disproportionately high rates of infection among Black, Asian and Hispanic children as a clear target of further study."

PEDSnet centers include Children's Hospital of Philadelphia; Cincinnati Children's Hospital Medical Center; Children's Hospital of Colorado; Nationwide Children's Hospital; Nemours Children's Health System; Seattle Children's Hospital; and, St. Louis Children's Hospital. Combined, these centers provide care to about 2.5 million children a year.

Highlights of their analysis include:

Like previous, smaller studies, this data shows that children are less likely to test positive and less likely to suffer severe illness when they do get infected.

Patients of African-American, Hispanic, and Asian race/ethnicity were less likely than white children to be tested. However, they were 2-4 times more likely to test positive.

Teens and young adults were more likely to test positive than younger children.

Children covered by Medicaid and other public programs were more likely to test positive than children from privately insured families.

Underlying cancer, diabetes (types 1 and 2), and other immune-suppressing conditions were indicators of increased risk of severe disease. But children with asthma were not found to be at increased risk of severe illness.

Among the 5,374 children who tested positive, 7% required hospital admissions. Of those hospitalized, 28% required intensive care and 9% required mechanical ventilation. Of the hospitalized children, eight died. (Case fatality rate: 0.2%)

"Further study is needed to understand the causes behind the variations in positivity rates," Pajor says. "How much is related to social determinants of risk, such as exposure to air pollution, housing density, or the likelihood of living with a person who must work at an in-person job? How much reflects differences in disease biology?"

Harnessing the Power of Big Data

The PEDSnet data coordinating center is based in Philadelphia, but the concept behind PEDSnet--launched in 2014--was a national cooperative effort among its co-founders, says Tracy Glauser, MD, Associate Director, Cincinnati Children's Research Foundation.

"Part of the challenge of pediatric research has been that many of our conditions are rare, so that no single institution has enough information by itself to comprehensively tackle certain issues," Glauser says. "The goal of PEDSnet has been to work out ways for institutions to share data to answer questions we cannot address alone."

Cincinnati Children's leaders have invested years of work into launching several data-sharing initiatives, including the Genomics Research and Innovation Network (GRIN) in 2015, becoming the data coordinating center for the Bench to Bassinet (B2B) Program for cardiac research in 2016 and being named the data coordinator for the Rare Diseases Clinical Research Network (RDCRN) in 2019.

Peter Margolis, MD, PhD, chaired the PCORnet Council and serves as the Cincinnati Children's site principal investigator for PEDSnet. Margolis is Co-Director of the James M. Anderson Center for Health System Excellence and has an extensive track record at building networks for health care quality improvement and research.

"PEDSnet provides a national digital architecture that can harness the power of the electronic health record to advance knowledge," Margolis says. "Without PEDSnet, gathering the information we are presenting today would have taken years."

MIS-C Needs Closer Study

While the latest study provides powerful data to address many questions, it also calls attention to the chaotic nature of the early days of the pandemic and how experts dealt with one of the most serious complications affecting children.

Early on children who experienced severe heart-damaging inflammatory reactions were diagnosed with Kawasaki disease, a very rare condition with largely unknown causes. As clinicians noted differences between the new cases and older ones, the diagnosis morphed into Kawasaki-like disease. It has since evolved into "multisystem inflammatory syndrome of childhood" (MIS-C).

The study co-authors say the medical community still has only a partial picture about the impact of MIS-C on children, in part because the rapidly evolving name changes has complicated data gathering. Also, with the pandemic still less than a year old in the U.S., much more study is needed to understand the long-term outcomes of MIS-C.

Much More to Learn

Some of the study's limitations include not including kids who were infected, or potentially killed, by COVID19 due to lack of testing availability. The study likely undercounts the actual numbers of asymptomatic infected children across the country, and does not address what risk those children may have presented to adults in their lives.

The participating medical centers overcame enormous technical challenges to build this tracking system early in the pandemic. Now, the data can be quickly refreshed to allow further, deeper analysis as the pandemic continues.

"Effective response to SARS-CoV-2 will require rapid but robust development of new clinical
and public health practices, based on a better understanding of viral and host biology," the co-authors say. "This knowledge will be critical not only in caring for severely ill patients, but also in constructing sustainable ways to minimize the disease burden caused by SARS-CoV-2."

The savanna tree frog Bokermannohyla ibitiguara is about 4 cm long and is found only in gallery forest along streams in the Serra da Canastra mountain range in the state of Minas Gerais, Southeast Brazil. In this watery forest environment, it can grow, feed, mate, and lay eggs without needing to range very far throughout its life cycle, according to a study published in Diversity and Distributions.

According to the Brazilian and US researchers who conducted the study, topography rather than vegetation is the main factor leading to more or less dispersal of the species in the territory, and this information is even recorded in its DNA.

They analyzed genetic variation among groups of B. ibitiguara living inside and outside the Serra da Canastra National Park, a protected area in the region, discovering that the flatter the terrain, the more genetically diverse is the population.

In areas of highly variable elevation, individuals are genetically similar. In evolutionary terms, this can be harmful to the species, which becomes more susceptible to disease and climate change, for example.

"Genetic analysis and conservation studies typically take land cover into account, among other factors, but the Cerrado [Brazilian savanna] is topographically diverse, including montane regions with high plateaus [chapadões] separated by low areas. We set out to verify whether this variable terrain played a part in the genetic diversity of the species, and found that it did. The vegetation alone didn't explain the genetic differences we identified between sites, or even within the same site. The topography did," said Renato Christensen Nali, first author of the article and a professor at the Federal University of Juiz de Fora's Institute of Biological Sciences (ICB-UFJF) in Minas Gerais, Brazil.

The study was one of the results of Nali's doctoral research, conducted at São Paulo State University's Bioscience Institute (IB-UNESP) in Rio Claro, Brazil, with a scholarship from FAPESP (São Paulo Research Foundation).

The research was part of the project "Reproductive ecology of anuran amphibians: an evolutionary perspective", for which the principal investigator is Cynthia Peralta de Almeida Prado, a co-author of the article. She is a professor at UNESP's School of Agrarian and Veterinary Sciences in Jaboticabal and teaches graduate students in zoology at IB-UNESP in Rio Claro.

The flatter the better

"The findings are very interesting because they bring to light a novel factor for conservation of the Cerrado, among other reasons. Ecological corridors and native forests are rightly considered important for conservation units, but more attention needs to be paid to the type of terrain. The topography should permit dispersal of the animals," said Nali, who heads ICB-UFJF's Amphibian Evolutionary Ecology Laboratory (Lecean).

To arrive at the results, the researchers analyzed 12 populations of B. ibitiguara, six inside Serra da Canastra National Park and six outside. Genetic diversity was much higher among the anurans living in the protected area than among those living outside the park. When the researchers correlated information on the degree of protection of the areas with the state of the vegetation, they found that these factors were less decisive for genetic diversity than the topography.

"The terrain is much more rugged outside the park, whereas inside it there's a large, very even plateau where the anurans can disperse more, find mates in more distant areas, and increase their genetic diversity," Nali said. "Outside the park, the rugged terrain and variable elevation appear to confine them to small areas."

The influence of these factors was evidenced by genetic tests. The researchers used a technique known as macrosatellite marker analysis to examine specific regions of the genome and found higher allele diversity in the populations living in the park. Allele diversity is one of the determinants of genetic integrity and adaptive potential.

In addition, the populations living outside the park displayed a greater loss of heterozygosity. If this loss, which is associated with declining genetic variability, recurs across several generations, it can eventually threaten the population's survival.

The study underscores the importance of topography as a factor to consider in conservation studies, as well as showing how the mere presence of a species in an area cannot ensure that it is not endangered.

"Molecular analysis enables us to find out if a population's genetic status is favorable," Nali said. "An area may have a large number of individuals, but DNA analysis may show that its genetic constitution is unfavorable, with few alleles and low heterozygosity. In practice, therefore, the population's effective size is small."

Although the study focused on only one species, he added, the findings can apply to others as well since the physical characteristics associated with dispersal are similar for other frogs and toads. More species need to be investigated to confirm the applicability of the findings.

The group noted that land cover nevertheless remains an important factor for conservation in the Cerrado, more than 50% of which has been converted into pasture or cropland, while less than 5% is protected by conservation units.

Deep neural networks, multilayered systems built to process images and other data through the use of mathematical modeling, are a cornerstone of artificial intelligence.

They are capable of seemingly sophisticated results, but they can also be fooled in ways that range from relatively harmless - misidentifying one animal as another - to potentially deadly if the network guiding a self-driving car misinterprets a stop sign as one indicating it is safe to proceed.

A philosopher with the University of Houston suggests in a paper published in Nature Machine Intelligence that common assumptions about the cause behind these supposed malfunctions may be mistaken, information that is crucial for evaluating the reliability of these networks.

As machine learning and other forms of artificial intelligence become more embedded in society, used in everything from automated teller machines to cybersecurity systems, Cameron Buckner, associate professor of philosophy at UH, said it is critical to understand the source of apparent failures caused by what researchers call "adversarial examples," when a deep neural network system misjudges images or other data when confronted with information outside the training inputs used to build the network. They're rare and are called "adversarial" because they are often created or discovered by another machine learning network - a sort of brinksmanship in the machine learning world between more sophisticated methods to create adversarial examples and more sophisticated methods to detect and avoid them.

"Some of these adversarial events could instead be artifacts, and we need to better know what they are in order to know how reliable these networks are," Buckner said.

In other words, the misfire could be caused by the interaction between what the network is asked to process and the actual patterns involved. That's not quite the same thing as being completely mistaken.

"Understanding the implications of adversarial examples requires exploring a third possibility: that at least some of these patterns are artifacts," Buckner wrote. " ... Thus, there are presently both costs in simply discarding these patterns and dangers in using them naively."

Adversarial events that cause these machine learning systems to make mistakes aren't necessarily caused by intentional malfeasance, but that's where the highest risk comes in.

"It means malicious actors could fool systems that rely on an otherwise reliable network," Buckner said. "That has security applications."

A security system based upon facial recognition technology could be hacked to allow a breach, for example, or decals could be placed on traffic signs that cause self-driving cars to misinterpret the sign, even though they appear harmless to the human observer.

Previous research has found that, counter to previous assumptions, there are some naturally occurring adversarial examples - times when a machine learning system misinterprets data through an unanticipated interaction rather than through an error in the data. They are rare and can be discovered only through the use of artificial intelligence.

But they are real, and Buckner said that suggests the need to rethink how researchers approach the anomalies, or artifacts.

These artifacts haven't been well understood; Buckner offers the analogy of a lens flare in a photograph - a phenomenon that isn't caused by a defect in the camera lens but is instead produced by the interaction of light with the camera.

The lens flare potentially offers useful information - the location of the sun, for example - if you know how to interpret it. That, he said, raises the question of whether adverse events in machine learning that are caused by an artifact also have useful information to offer.

Equally important, Buckner said, is that this new way of thinking about the way in which artifacts can affect deep neural networks suggests a misreading by the network shouldn't be automatically considered evidence that deep learning isn't valid.

"Some of these adversarial events could be artifacts," he said. "We have to know what these artifacts are so we can know how reliable the networks are."

While respiratory issues continue to be the most common symptom of a COVID-19 infection, new research indicates the disease could also be associated with hypercoagulability, or increased tendency of the blood to clot. In a new study published November 20, 2020 in the journal EClinical Medicine by The Lancet, researchers from UC San Diego Health found that blood clots led to an increased risk of death by 74 percent.

Led by Mahmoud Malas, MD, division chief of Vascular and Endovascular Surgery at UC San Diego Health, researchers reviewed 42 different studies involving more than 8,000 patients diagnosed with COVID-19. Using random models, the team produced summary rates and odds ratios of mortality in COVID-19 patients with thromboembolism, blood clots -- and compared them to patients without these conditions to determine what effect blood clots may have on risk of death.

"We began to notice a really unusual manifestation of venous and arterial thromboembolism in patients with COVID-19," said Malas. "In addition to higher instances of blood clots, the mortality for patients hospitalized for COVID-19 and with thromboembolism was much higher, compared to patients without clots. It's unusual because we have never seen anything like this with other respiratory infections."

Overall, 20 percent of the COVID-19 patients were found to have blood clots in the veins, and among patients in the intensive care unit, that statistic increased to 31 percent.

Blood clots in the vein, or deep vein thrombosis, can reach the lungs and develop into pulmonary embolism, resulting in higher risk of death. Blood clots in the arteries may lead to limb amputation if not treated surgically in a timely fashion.

In the study, Malas and colleagues performed a systemic review through meta-analysis, which is a statistical method that allowed researchers to combine multiple studies to produce a single comprehensive paper.

"The collective experience in the literature as captured in this meta-analysis study brings additional light on the importance of blood vessel clotting events in hospitalized patients with COVID-19," said Bryan Clary, MD, surgeon-in-chief at UC San Diego Health and co-author of the study. "While the frequency of these events is much higher than expected, our study likely underestimates the incidence of thromboembolism in the global population of patients with COVID-19, including non-hospitalized patients."

According to Malas, arterial blood clots developing in people with the flu is extremely rare, and the rate of clotting in patients with COVID-19 is higher than what is reported for other viral pandemics, including the H1N1 influenza of 2009.

Similar symptoms are shared between influenza and SARS-CoV-2, such as fever, cough, shortness of breath, or fatigue. Blood clotting can occur in patients hospitalized with the flu, but only in veins. For patients with COVID-19, blood clots can appear in either veins or arteries.

Typically, clotting in the arteries is caused by health factors, such as atrial fibrillation, high blood pressure, high cholesterol, diabetes, or lifestyle choices like smoking. Patients who are hospitalized for long periods of time are also more at risk for blood clots in the vein due to immobility.

Blood clots in the vein are treated or prevented with prescribed blood thinners. Proactively administering such medications to hospitalized patients can also help prevent clots from forming. Clinical trials are ongoing to determine how blood thinners can reduce the risk of clotting in patients with COVID-19.

"What we can learn from this paper is due diligence," said Malas. "We're still in the process of understanding the pathophysiology of COVID-19, so it's important to have a low index of suspicion when it comes to this infection to ensure we're doing all we can to mitigate the spread and prevent severe outcomes."