A Novel Approach to Creating Tailored Odors and Fragrances Using Machine Learning

Can we use machine learning methods to predict the sensing data of odor mixtures and design new smells? A new study by researchers from Tokyo Tech does just that. The novel method is bound to have applications in the food, health, beauty, and wellness industries, where odors and fragrances are of keen interest.

The sense of smell is one of the basic senses of animal species. It is critical to finding food, realizing attraction, and sensing danger. Humans detect smells, or odorants, with olfactory receptors expressed in olfactory nerve cells. These olfactory impressions of odorants on nerve cells are associated with their molecular features and physicochemical properties. This makes it possible to tailor odors to create an intended odor impression. Current methods only predict olfactory impressions from the physicochemical features of odorants. But that method cannot predict the sensing data, which is indispensable for creating smells.

To tackle this issue, scientists from Tokyo Institute of Technology (Tokyo Tech) have employed the innovative strategy of solving the inverse problem. Instead of predicting the smell from molecular data, this method predicts molecular features based on the odor impression. This is achieved using standard mass spectrum data and machine learning (ML) models. "We used a machine-learning-based odor predictive model that we had previously developed to obtain the odor impression. Then we predicted the mass spectrum from odor impression inversely based on the previously developed forward model," explains Professor Takamichi Nakamoto, the leader of the research effort by Tokyo Tech. The findings have been published in PLoS One.

Researchers devise tunable conducting edge

In their experiments, the researchers stacked monolayer WTe2 with Cr2Ge2Te6, or
CGT. Credit: Shi lab/UC Riverside

A research team led by a physicist at the University of California, Riverside, has demonstrated a new magnetized state in a monolayer of tungsten ditelluride, or WTe2, a new quantum material. Called a magnetized or ferromagnetic quantum spin Hall insulator, this material of one-atom thickness has an insulating interior but a conducting edge, which has important implications for controlling electron flow in nanodevices.

In a typical conductor, electrical current flows evenly everywhere. Insulators, on the other hand, do not readily conduct electricity. Ordinarily, monolayer WTe2 is a special insulator with a conducting edge; magnetizing bestows upon it more unusual properties.

“We stacked monolayer WTe2 with an insulating ferromagnet of several atomic layer thickness — of Cr2Ge2Te6, or simply CGT — and found that the WTe2 had developed ferromagnetism with a conducting edge,” said Jing Shi, a distinguished professor of physics and astronomy at UCR, who led the study. “The edge flow of the electrons is unidirectional and can be made to switch directions with the use of an external magnetic field.”

Shi explained that when only the edge conducts electricity, the size of the interior of the material is inconsequential, allowing electronic devices that use such materials to be made smaller — indeed, nearly as small as the conducting edge. Because devices using this material would consume less power and dissipate less energy, they could be made more energy efficient. Batteries using this technology, for example, would last longer.

Bronchodilators don’t improve smoking-related respiratory symptoms in people without COPD

Credit: Ralph from Pixabay

Researchers supported by the National Institutes of Health have found that dual bronchodilators – long-lasting inhalers that relax the airways and make it easier to breathe – do little to help people who do not have chronic obstructive pulmonary disease (COPD), but who do have respiratory symptoms and a history of smoking.

COPD, a lung disease that obstructs the airways and leads to coughing, wheezing, and shortness of breath, affects about 15 million Americans. However, millions of others who smoke or used to smoke and have some symptoms of COPD have also been prescribed bronchodilators.

“We’ve assumed these medications worked in patients who don’t meet lung function criteria for COPD, but we never checked,” said MeiLan K. Han, M.D., a principal investigator and first author of the study. “We now know these existing medications don’t work for these patients.”

The findings of the study, which was funded by the National Heart, Lung, and Blood Institute (NHLBI), were published in the New England Journal of Medicine and simultaneously presented at the European Respiratory Society International Congress.

According to scientists, the implications are significant. First, they show the importance of diagnosing lung conditions through spirometry, a lung function test Han noted is underutilized in clinical practice. Second, they show the need for new, effective therapies for patients without COPD.

New treatment could result in more donor lungs

Sandra Lindstedt, Snejana Hyllén, and Leif Pierre
Credit: Lund University

A large number of lungs donated cannot be used for transplantation. Researchers at Lund University in Sweden and Skåne University Hospital have conducted an animal study bringing hope that more donor lungs could be used in the future. Researchers have launched a pilot study to investigate whether the treatment will have the same positive effects on human beings.

About 190 organs are donated in Sweden every year. Due to injuries to the lungs, only about 30 percent of them can be used for transplantation. Adding to that, the mortality rate is high: about half of the patients pass away within five years of transplant.

” The results from our study indicate that a certain treatment can help us use a larger part of a donor lung, and that there is an improved outcome during the first two days after surgery”, says Sandra Lindstedt, senior consultant in thoracic surgery at Skåne University Hospital and adjunct professor at Lund University.

In their study on pigs, the researchers investigated the effects of reducing the levels of cytokines in lungs. Cytokines are small proteins that are produced by specific cells of the immune system.

High plant diversity is often found in the smallest of areas

This meadow in Romania is one of the most species-rich regions on earth - in 2009, a research team found 98 plant species here.
Credit: Jürgen Dengler

It might sound weird, but it's true: the steppes of Eastern Europe are home to a similar number of plant species as the regions of the Amazon rainforest. However, this is only apparent when species are counted in small sampling areas, rather than hectares of land. An international team of researchers led by the Martin Luther University Halle-Wittenberg (MLU) and the German Centre for Integrative Biodiversity Research (iDiv) Halle-Jena-Leipzig has now shown how much estimates of plant diversity change when the sampling area ranges from a few square meters to hectares. Their results have been published in the journal "Nature Communications" and could be utilized in new, more tailored nature conservation concepts.

In their study, the team analyzed a dataset of around 170,000 vegetation plots from all of the Earth’s climate zones. The data included information on all of the plant species found at a location and the coordinates of the respective area under study. The data was taken from the globally unique vegetation database "sPlot", which is located at iDiv.

"Most studies on global biodiversity are conducted on a relatively large scale, for example at a state or provincial scale. We wanted to find out how much results differ when smaller areas are examined," says Professor Helge Bruelheide from MLU. The team used artificial intelligence to investigate, among other things, the relationship between the number of plant species and the size of the area under study.

Boy’s discovery reveals highly complex plant-insect interaction

Researchers discovered that not only do gall wasps manipulate oaks to produce galls, but they also manipulate ants to retrieve the galls to their nests, where the wasp larvae may be protected from gall predators or receive other benefits. 

Credit: Michael Tribone, Penn State.

When 8-year-old Hugo Deans discovered a handful of BB-sized objects lying near an ant nest beneath a log in his backyard, he thought they were a type of seed. His father, Andrew Deans, professor of entomology at Penn State, however, knew immediately what they were — oak galls, or plant growths triggered by insects. What he didn’t realize right away was that the galls were part of an elaborate relationship among ants, wasps and oak trees, the discovery of which would turn a century of knowledge about plant-insect interactions on its head.

Looking back, Hugo, now 10, said, “I thought they were seeds, and I felt excited because I didn’t know ants collected seeds. I always thought ants would eat food scraps and stuff around the house. Then I got more excited when [my dad] told me they were galls, because [my dad] was so excited. I was surprised that ants would collect galls because why would they do that?”

According to Andrew Deans, who is also the director of Penn State’s Frost Entomological Museum, many plant-insect interactions are well documented. For example, most "cynipid" wasp species have long been known to induce oak trees to produce protective galls — or growths — around their larvae to ensure the safety of their developing offspring. Additionally, certain plants — including bloodroot (Sanguinaria canadensis), a wildflower native to North America — produce edible appendages, called elaiosomes, on their seeds to attract ants, which then disperse the seeds by carrying them back to their nests. This latter example is referred to as "myrmecochory" — or seed dispersal by ants.

How Artificial Intelligence can explain its decisions

They have brought together the seemingly incompatible inductive approach of machine learning with deductive logic: Stephanie Schörner, Axel Mosig and David Schuhmacher (from left).
Credit: RUB, Marquard

If an algorithm in a tissue sample makes up a tumor, it does not yet reveal how it came to this result. It is not very trustworthy. Bochum researchers are therefore taking a new approach.

Artificial intelligence (AI) can be trained to recognize whether a tissue image contains tumor. How she makes her decision has so far remained hidden. A team from the Research Center for Protein Diagnostics, or PRODI for short, at the Ruhr University Bochum is developing a new approach: with it, the decision of an AI can be explained and thus trustworthy. The researchers around Prof. Dr. Axel Mosig in the journal "Medical Image Analysis".

Bioinformatician Axel Mosig cooperated with Prof. Dr. Andrea Tannapfel, head of the Institute of Pathology, the oncologist Prof. Dr. Anke Reinacher-Schick from St. Josef Hospital of the Ruhr University as well as the biophysicist and PRODI founding director Prof. Dr. Klaus Gerwert. The group developed a neural network, i.e. an AI that can classify whether a tissue sample contains tumor or not. To do this, they fed the AI with many microscopic tissue images, some of which contained tumors, others were tumor-free.

Scientists discover new ant species

Three-dimensional image of the previously unknown extinct ant species.
Illustration credit: Hammel/Lauströer

An international team of scientists has discovered a previously unknown extinct ant species encased in a unique piece of amber from Africa. Using the X-ray light source PETRA III at the German Electron Synchrotron (DESY) in Hamburg the researchers, from Friedrich Schiller University Jena, the University of Rennes in France, the University of Gdansk in Poland, as well as the Helmholtz-Zentrum Hereon in Geesthacht, Germany, had examined the critical fossil remains from 13 individual animals in the amber and realized that they could not be attributed to any previously known species. The name given to the new species and genus is †Desyopone hereon gen. et sp. nov. In this way, the scientists are honoring the two research institutions involved – DESY and Hereon – which contributed significantly to this find with the help of modern imaging techniques. Ultimately, it was only possible to identify the new species and genus through the combination of extensive phenotype data from scans and recent findings from genome analyses of living ants. The team reports on its discovery in the research journal “Insects”.

These mice grow bigger on the rainier sides of mountains. It might be a new rule of nature

Shaggy soft-haired mouse Abrothrix hirta (Order Rodentia, Family Cricetidae)
Resized Image using AI by SFLORG
Credit: Pablo Teta

Scientists studying mice from the Andes Mountains in Patagonia noticed something they couldn’t explain: the mice from the western side of the mountains were bigger than the ones from the east, but DNA said that they were all from the same species. The researchers examined the skulls of 450 mice from the southern tip of South America, and found that existing biological laws didn’t explain the size differences. Instead, in a new paper in the Journal of Biogeography, the scientists put forth a new hypothesis: the mice on the western slopes were bigger because that side of the mountain range gets more rain, which means there’s more plentiful food for the mice to eat.

“There are a bunch of ecogeographic rules that scientists use to explain trends that we see again and again in nature,” says Noé de la Sancha, a research associate at Chicago’s Field Museum, an assistant professor of Environmental Science and Studies at DePaul University, and the paper’s corresponding author. “With this paper, I think we might have found a new one: the rain shadow effect can cause changes of size and shape in mammals.”

The mice that de la Sancha and his colleagues examined in this study are shaggy soft-haired mice, Abrothrix hirta. “They’re very cute little buggers, they have soft white bellies,” says de la Sancha. “They live in the mountains, which makes them unique, but they’re also found in lower elevations. Overall, they’re not very well-studied.”

New methodology predicts coronavirus and other infectious disease threats to wildlife

The rate that emerging wildlife diseases infect humans has steadily increased over the last three decades. Viruses, such as the global coronavirus pandemic and recent monkeypox outbreak, have heightened the urgent need for disease ecology tools to forecast when and where disease outbreaks are likely. A University of South Florida assistant professor helped develop a methodology that will do just that – predict disease transmission from wildlife to humans, from one wildlife species to another and determine who is at risk of infection.

The methodology is a machine-learning approach that identifies the influence of variables, such as location and climate, on known pathogens. Using only small amounts of information, the system is able to identify community hot spots at risk of infection on both global and local scales.

“Our main goal is to develop this tool for preventive measures,” said co-principal investigator Diego Santiago-Alarcon, assistant professor of integrative biology. “It’s difficult to have an all-purpose methodology that can be used to predict infections across all the diverse parasite systems, but with this research, we contribute to achieving that goal.”

With help from researchers at the Universiad Veracruzana and Instituto de Ecologia, located in Mexico, Santiago-Alarcon examined three host-pathogen systems – avian malaria, birds with West Nile virus and bats with coronavirus – to test the reliability and accuracy of the models generated by the methodology.