Welcome to the October edition of STEM Spectrum Monthly News, presented to you by The STEM Spectrum's News Editors! Each month, the latest advancements in science, technology, engineering, and mathematics are broken down and analyzed. This month's edition zooms in on various topics, from AI-based drug repurposing to climate change and extreme weather.

Can Artificial Intelligence now repurpose Physical Medicine?

Affecting over 300 million people worldwide, there are more than 7000 undiagnosed, rare diseases running worldwide. Most of these, therefore, remain untreated or undertreated with only 5-7% having an FDA-approved drug.

Artificial Intelligence is a revolutionary discovery and invention paving its way through human lifestyle. Present in growing avenues of daily life: from school lessons to grocery store cashiers to automobiles - one avenue in which AI was perceived as untouchable has been the medical field. This perception, however, is falling shorter and shorter as researchers find new ways to integrate the technology within healthcare practices.

Specifically, the AI model known as TxGNN, the first one developed to identify drug candidates for rare diseases with no treatments, has been able to detect drug candidates from existing medicines for more than 17,000 diseases. The model has, therefore, found the largest number of diseases any AI model can handle.

TxGNN investigations and trials were led by scientists at Harvard Medical School. Marinka Zitnik, assistant professor of biomedical informatics at HMS, stated the institution aims to “identify new therapies across the disease spectrum;” as they foresee this new model to be capable of narrowing a gap that is currently creating serious health disparities. Such AI models prove promising to reduce the global disease burden - making it easier, faster and more cost-effective to develop innovative therapies with existing drugs as it doesn’t rely on the need of designing new drugs from scratch.

TxGNN has two main features: one identifying candidates plausible for treatment and another which explains the rationale for the decisions. The tool identified candidates from over 800 FDA-approved and experiential medicines whilst simultaneously predicting which drugs would impose side effects for specific conditions. This proves greatly advantageous when compared with the standard method in practice currently, where trial and error is required to find contraindications.

Using already approved drugs proves an alluring way to develop treatments as it has already been studied and tested for safety. Thus, repurposed drugs can offer alternatives with fewer side effects and more guaranteed returns.

This increases the capacity and expands limits of medicine as it engages clinicians to find better, improved ways of recycling knowledge they already have. This not only holds a higher success rate but is a way for the medical field to be more sustainable and efficient in the care they provide. With this unprecedented capacity, TxGNN can expedite drug repurposing unlike ever before - finding causes, treatments and solutions for the health crisis thousands are suffering under.

The Growing Patriarchy in Physician Divide Due To Mental Health Struggles

The divide between men and women has been present for centuries in almost every aspect of life, whether it be job pay, job opportunities. A recent study by Eva Schernhammer, an epidemiologist at Harvard Chan School, shows how this is now prevalent in the mental health of physicians, where women physicians have a 24% higher risk of suicide in comparison to the general population.

While this may be shocking, it is a stark improvement from 20 years ago when this statistic was more than 3 times larger - giving female physicians 76% higher odds. It is noted that, in general, physicians have holistically experienced higher rates of suicide than the general public - mostly due to the arduous pressure of their working environment. Male physicians have long been said to have higher risk, but this had not been thoroughly evaluated nor compared with females because of the lack of data for female physicians until the early 2000s.

Today, females make up more than 50% of medical students in the U.S. alone. The fact that female suicide rates are concerning and need to endorse caution for the mental care of potential sufferers. When discussing potential factors that could have been the root cause contributing to higher suicide rates among female physicians, colleagues pointed to a double burden as females must balance their work along with the household. Despite these applying to both men and women, Schernhammer proposes a higher demand strained on women.

Secondly, 20 years ago, sexual harassment was a big contender in denoting the reasoning for such a high statistic. Another report showcased that three-quarters of all women experienced some form of harassment from male colleagues within their working environment. As such, there could be a direct correlation with that and the impeding stress on their mental health.

The last component analysis conducted by the researcher focused on a more specific aspect, she compared the rate of attempts with actual suicides and found that women physicians have a much higher completion rate than women in the general population. What’s more, they also tend to use different means of committing - finding higher likeliness to use poisoning since they are equipped with more knowledge on what can effectively pull the trigger.

All data studied, there are some limitations to the study. Primarily, geographical evolution and constraints hinder the replicability of such statistics across the world. Schernhammer notes how the demographic in the US, which has modernized over the last 20 years, differs starkly from lesser developed countries, like India. She states: “We just heard more recently how physicians are being attacked and harassed, particularly female physicians.” The researcher believes that with more equal rights for women, where these issues are talked about more, and a more welcoming environment is created for females, it could support alleviating some of the risk.

Schernhammer offers some probable solutions to attempt paving a way for improvement in these statistics. She understand that major shifts have been done from 20 years ago till now, contributing to the declining statistics and she notes how work hour reductions - thanks to studies from Chuck Czeisler at Harvard Medical School which highlighted the increased number of medical errors from lack of sleep - can be one of the first steps to reducing the mental burden on doctors.

Overall, Schernhammer hopes her research can offer clarity and take away feelings of fighting this way alone. She desires the creation of more awareness so people feel it as less of a stigma if they are suffering from it. She wants people to know they are not the only ones and that help is an option - an option that is available to them. For males or females, physicians or stay-at-home individuals; your mental health is not just a number but a priority.

Climate Change: A Disaster Causing More Disasters

Climate change has become a topic of conversation in nearly every community. It refers to the long-term shifts in temperature and weather patterns which were once simply caused by natural phenomena, but are now more largely influenced by anthropogenic causes.

Recently, one of the strongest hurricanes of this year, Hurricane Milton explosively intensified from Category 1 to Category 5 in just 20 hours with winds of 290 kilometers per hour. Meteorologists have watched in awe and horror as this hurricane has transformed into one of the strongest Atlantic storms ever recorded. These rapid 20-hour intensifications of storms Hurricane Milton and supporting Hurricane Helene were primarily fueled by the Gulf’s disproportionately warm water. Tropical storms can suck up heat from warm seawater, dragging the humid air upwards where it will condense and release heat into the storm’s core. As the storm moves forward, more water and heat are pumped into the ai, causing the spiraling winds to move faster and with greater force.

Consequently, such warm gulf waters were predicted as to be derived by human-caused climate change. Gulf of Mexico sea surface temperatures were reported as being 1.26 degrees Celsius warmer than they would have normally been without climate change. What this means in hurricane terms is that the extremely high temperatures along Helene’s path were made 200-500 times more likely simply due to climate change.

Specifically, an International World Weather Attribution analysis, published on October 9, highlighted the role of climate change in being a big contributor to the intensification of Hurricane Helene. Ielene poured 50-75 centimeters of rain, causing major flooding and hundreds of deaths across the Southeast US. Such heavy rainfall was determined as 10% heavier than if human-caused climate change was not as potent as it is now.

Such sea surface temperatures and rainfall analysis promote the major message that climate change is happening right now, and is inducing graver, stronger impacts than before. It was a primary cause of both deadly storms, and we, as humans, are here to blame for such events surging to such a high extent in such a short amount of time. Climate scientist, Daniel Gilford, urges people to “sit up and take notice.” “We are to blame for these storms and ironically we are to blame for the damage these storms have inflicted on us.” By harming our environment, it truly won’t be long before the world around us begins to fight back, take control, and cause more harm than we could imagine.

'Islands' of regularity discovered in the famously chaotic three-body problem

One of the most famous problems in physics and astronomy is the three-body - problem. When three massive objects meet in outer space, their gravitational pull influences each one in an unpredictable way; it is chaotic by nature. New research by a researcher at the University of Copenhagen has discovered that oftentimes in these situations the system avoids chaos and resorts to regular patterns, by expelling one of the bodies of the system.

This problem has been around for centuries, with physicist Isaac Newton first describing this not just complex but chaotic problem. The current theory states that when three bodies meet, their interaction evolves chaotically. This means they do not follow regular patterns, but move unpredictably, we do not know how the system will move next. However, amid this chaos there exists some gaps where the bodies seem to move in predictable patterns which depend on the body's position, speed, and angle of approach.

Researchers have been working on simulations for years to understand the three-body problem and if there is a way to solve it. Doing so can bring about great insights to the topic of Gravitational waves and help uncover mysteries of gravity, and other mysteries of the universe, including black holes.

The researcher from Copenhagen developed his software program known as Tsunami, which calculates the movements of these bodies based on the laws of physics. He ran millions of simulations with various parameters and he uncovered that at some points given specific conditions there were some predictability with the system. In his findings, oftentimes the object with the lowest mass would eventually be ejected from the solar system, thus causing a regular predictable system afterwards.

This discovery was crucial to understanding and solving the problem. If the system was a pure chaos system, we can make accurate calculations using statistical methods, but because the chaos is interwoven by periods of regularities, the calculations become extremely complex. It throws off our traditional statistical calculations, which is why it is difficult to predict the motion of these bodies. However, this discovery has enabled us to realize that through modified statistical calculations and numerical analysis it may be possible to accurately predict the motion of the bodies, solving the three-body problem.

New paradigm of drug discovery with world's first atomic editing

One of the most significant difficulties when it comes to treating illnesses and diseasess is developing drugs and medicines with high efficacy. Over 12,500 drugs are being developed in the United States alone, but oftentimes it takes years to make significant progress and have them available to the masses. However, researchers at the Korea Advanced Institute of Science & Technology have made a groundbreaking advancement in drug development by creating technology that allows for the precise editing of key atoms in drug molecules which enhances their effectiveness.

This technology and atom editing technique were developed by Professor Yoonsu Park’s team and are the first of their kind in the world. Specifically, the team announced that they developed a process to convert Oxygen atoms in furan compounds, which are an organic compound composed of Carbon and Oxygen, into nitrogen atoms. These nitrogen atoms had a specific shape and form that is commonly used in developing pharmaceuticals.

In many pharmaceutical products and medicines, the effectiveness of a drug depends on a single atom, usually oxygen or nitrogen, and these are what impact how well the drug works. By editing the atoms within a drug of another type, we can convert one type to another, one that will boost the efficacy of the drug developed. This enables people to see far more beneficial outcomes and the drug can be used to a vastly increased number of people.

The specific technique they used was that they used a photocatalyst, which is a molecule that uses light energy to cut and replace atoms. This allowed them to make precise edits to drug molecules at room temperature and enabled them to remove oxygen atoms and add in nitrogen atoms instead. This replacement increased the effectiveness of the drug as Nitrogen often boosts how well the drug’s effect works on any particular illness.

This breakthrough could revolutionize how drugs are designed and made, enabling scientists to build drugs that are far more effective. Not only that, but it can significantly lower the cost of drugs by expediting the development by making more effective versions in a relatively efficient manner. This discovery has received national and international recognition, and this discovery can and will lead to the development of new, and more effective treatment for diseases in less time.

AI speeds up the discovery of energy and quantum materials

Researchers from the Massachusetts Institute of Technology and Tohoku University have developed an AI tool that can predict the optical properties of materials as accurately as quantum simulation but millions of times faster. This tool can enable accelerated development of photovoltaic and quantum materials used in a plethora of domains.

Innovating technologies such as LEDs, solar cells, photodetectors and microprocessors require extensive knowledge of the optical properties of the materials they are made of. This takes an extreme amount of time and requires complex mathematical equations and computational power. Quantum computers, which are currently used to simulate these processes, often also struggle to compute the necessary information to uncover the properties of these materials. However, these researchers have developed an AI tool that can predict these properties even faster, which can allow for the accelerated development of all of these minerals we use in our daily lives.

The traditional way of calculating optical properties, which involves complex mathematical equations and computationalpowere, makes it extremely slow and expensive to test and develop these materials. The AI model that was created by these researchers overcomes this challenge. The way it works is that it uses the material's crystal structure at the atomic level to predict the behavior it has with optical light at various frequencies. These predictions can then be tested by the computers and the researchers themselves to verify which expedite the process significantly.

The model leverages machine learning to specifically graph neural networks to represent the molecules and materials in a simple yet efficient way. The researcher developed a new algorithm for the model known as “universal ensemble embedding”, which improves the accuracy of its predictions and can be applied to any neural network model, assisting millions of other AI models across the world.

The optical properties of materials are extremely important as they play a crucial role in analyzing whether a material is suitable to be used for the implementation of LEDs and solar cells, and cheaper materials can allow for cheaper products that can boast the usage of these products. The problem with finding these optical properties experimentally is that it requires precise lasers to accomplish, but most lasers have limitations that simply inhibittheirs usage. Most researchers resort to using computers, but the simulations are so complex that the computational cost incurred is detrimental.

Thus, this machine learning model has played a crucial role in identifying optical properties but also revolutionizing this field. The method enables highly precise prediction and can be used for a plethora of other applications in the scientific realm. Going forward, the scientist aims to use the model to create databases of the properties of materials and hopes to improve the model’s ability and discover materials that can be used to create a plethora of products that support mankind.

Presented to you by the TSS News Editing Team.

October STEM News