I’M SITTING ON my bed in sweatpants and a T-shirt—feet up, legs crossed—reaching out to a potential source via Twitter DMs.
He wants to talk through Google Hangouts, but Skype is a better option for me for recording since I’m already set up for it. What is also going through my mind is whether my dog, Oreo, will start barking during said interview, but then I realize there are just some variables that are out of our control.
At that very moment, I think to myself: What a wonderful, technologically advanced world.
Like millions of people around the globe, I’ve been working from home the past few weeks because of the COVID-19 pandemic, and I don’t know when I’ll be able to return to the office. But this isn’t about me or them. This is about us tackling a problem together. During a crisis, technology and innovation become the focal points for not only benefiting society now but also creating a better future in the process.
While there are many unanswered questions at this point, scientists, innovators, governments and companies around the world are on it as I type these words. The speed at which change is coming to tackle this enormous problem is nothing short of amazing and fascinating to watch. With all that’s at stake, a barking dog is the least of our concerns.
The speed at which COVID-19 is being researched and analyzed is astounding and should offer a great deal of hope, especially when one looks back at one of the worst viral outbreaks in recorded history: the 1918 influenza pandemic. When that virus hit and hit again and hit again, it’s fair to say no one knew what hit them. As the U.S. Centers for Disease Control and Prevention details on its website, the first wave reached the United States in March 1918 followed by a second wave in September 1918 and a third one in spring 1919. Overall, the death toll was an unfathomable 50 million people worldwide. Of course, there were no vaccines at the time. People wore masks (albeit primitive ones, which were nothing like today’s N95 masks that are designed to filter out 95% of airborne particles, including viruses) and businesses and schools were closed. It might sound familiar, but technology has obviously improved phenomenally since then.
When this article was first published in April 2020, the total confirmed deaths from COVID-19 in the United States was 45,500. For perspective, when we started writing this article earlier in the year, the number was 11,252, but as of Nov. 2, 2020, the number of deaths has risen to an unfathomable 231,000 in the United States and 1.2 million worldwide with no imminent end to the virus in sight.
While the number of deaths is scary and will undoubtedly get worse before it gets better, there is a worldwide effort going on right now to “flatten the curve” and limit the number of people who are affected by this disease through social distancing. Without taking measured steps, the “curve” of cases will continue to rise before slowly coming down, but by staying home and limiting the spread of the virus, the “curve” is flattended considerably. Technology plays a huge role in that. At the most basic level, many workers can work from home in a way that was not possible even, say, 10 years ago.
Besides remote work, and you’ve heard this over and over, proper hand-washing, not touching your face and social distancing are all important tools to help stop the spread.
“Social distancing will work,” said Stefan Green, the associate director of the Research Resources Center and director of the Genome Research Core at the University of Illinois at Chicago. “The critical thing is to decrease the rate of transmission to gain time to develop a vaccine.”
Unlike previous pandemics including the 1918 flu: “Nobody knew about DNA or RNA, and there was no chance that you had a sequence of the entire organism,” Green said.
“And here, within two weeks, we were able to identify what we’ve got. So our ammunition in terms of scientific and technical capabilities are much higher than they were 100 years ago, and I think that should all give us great hope that we can properly deploy our societal resources to address this pandemic.”
In other words: Technology is the differentiator.
The mystery of the 1918 flu transmission took decades and separate trips to Brevig Mission, Alaska, to solve. There are 72 Iñupiat Eskimos interned there near Port Clarence on the far western side of the Last Frontier. There were only 80 people living in the village at the time.
The corpses were mostly preserved from the permafrost, so the theory was that the frozen bodies of those unfortunate souls could help solve the mystery of what exactly caused the 1918 outbreak. Yet, it would take two trips separated by 46 years for microbiologist Johan Hultin to find the lung tissue sample he needed to solve the mystery. Twice he ventured to Brevig Mission—in 1951 and 1997—to collect samples from the deceased Eskimos there, and twice he asked for and received permission from village elders to dig into the grave to remove tissue to research the disease.
During the first trip, Hultin built a fire to warm the ground, and he then shoveled through the melted soil and exposed the surface to air just to get to the mass grave below—a painstaking operation that took “two days to reach the first body,” Hultin told the Anchorage Daily Tribune in 2014. Even with the preserved tissues, Hultin and his research team were not successful in reviving the virus for research.
The second trip, which Hultin again volunteered to take, but this time at the age of 72, led to a sample that could be reconstructed and researched. The virus, it was discovered, originated in birds.
That 1918 jump-rope song was more on-point than people at the time could have ever imagined.
While there’s still so much that isn’t known about COVID-19, scientists and innovators are working feverishly to try to stop the spread of the virus, which likely came from bats via a “spillover,” meaning it jumped from animals to humans. Although,one article suggests the virus could be a “chimera” of two different viruses, possibly originating in bats, pangolins or even another species. Either way, it’s clear the virus is zoonotic. So let’s start there.
Researchers like Shi Zhengli, a principal investigator at the Wuhan Institute of Virology who is known as “Bat Woman” in China, go into caves to collect samples of blood, saliva and guano from a colony of bats to learn more about the many diseases those nocturnal creatures carry, including corona viruses. As Scientific American explains, Shi first got a call about a novel corona virus on Dec. 30, 2019. Shortly after, she told the publication, “Using a technique called polymerase chain reaction, which can detect a virus by amplifying its genetic material, the first round of tests showed that samples from five of seven patients contained genetic sequences known to be present in all coronaviruses.”
Meanwhile, halfway around the world a Canadian artificial intelligence company named BlueDot was tracking a cluster of “unusual pneumonia” cases in Wuhan. But, Kamran Khan, founder and CEO of BlueDot, told CNBC, “We didn’t know at that moment that this was going to become something of this magnitude.” The original idea for BlueDot, he explained in the article, was to “spread knowledge faster than the diseases spread themselves” and to ensure people don’t get caught “fl at-footed” during a new outbreak. The fast-moving pace of this virus fit the bill perfectly.
On Jan. 7, 2020, just over a week after Shi got her call and BlueDot started tracking the cluster, Shi and her team—based on results from the polymerase chain reaction (PCR), genome sequencing and antibody testing—learned that the virus, officially called SARS-CoV‑2, “was 96% identical to that of a corona virus the researchers had identified in horseshoe bats in Yunnan.”
Two days later, the WHO made its first statement about the cluster in Wuhan, China, noting: “Preliminary identification of a novel virus in a short period of time is a notable achievement and demonstrates China’s increased capacity to manage new outbreaks.” The statement went on to say: “WHO does not recommend any specific measures for travelers. WHO advises against the application of any travel or trade restrictions on China based on the information currently available.”
On Jan. 30, a little over three weeks after that statement, a WHO Emergency Committee determined that COVID-19 met the criteria for a Public Health Emergency of International Concern (PHEIC), and the WHO director-general made it official. The same day, the U.S. State Department raised its travel advisory to Level 4, meaning “Do not travel” to China.
A dramatic development in a short period of time—and the world was just beginning to understand the sheer magnitude of this situation.
THE NEED FOR SPEED
While the human brain is an amazing organism that can store a petabyte worth of memories, as we explained in our fall/winter 2019 cover story, there are still issues with being able to access the information and analyze it. Artificial intelligence and machine learning on the other hand can provide useful analytics much quicker, perhaps in seconds, so when Chinese researchers posted the sequence for COVID-19 in a public database on Jan. 10, 2020, the race began to find a vaccine—a race that could take a year or longer.
In July 2019, a team of researchers from Flinders University in South Australia released the first flu vaccine created solely by artificial intelligence via a program called SAM, which is short for Search Algorithm for Ligands.
“SAM is based on a deep machine learning algorithm,” Nikolai Petrovsky told Spark in an email. Petrovsky, who led the research into the AI vaccine, is a professor at Flinders University in the College of Medicine and Public Health. “We essentially train SAM on known compounds with particular features such as anti-COVID-19 antiviral activity, and then SAM can use the information learned to look at large chemical libraries containing millions of compounds and identify compounds SAM thinks might have antiviral activity. We can then make these compounds and test them against COVID-19. So in this case SAM can be used not just to help design vaccines, as we did previously, but also to help find new antiviral drugs.”
Petrovsky said that screening those millions of compounds takes mega-computational power, which SAM gets from the cloud. It “is not something we could have easily done even a few years ago,” he told Spark, adding: “The speed with which information can be gleaned in this way is unprecedented, and speed is what is needed when dealing with a potential pandemic virus as we have found with COVID-19.”
About 10,000 miles away, researchers at Oak Ridge (Tennessee) National Laboratory have tapped into Summit, a supercomputer that can make 200 trillion calculations per second. For perspective, while there’s no exact answer, it would probably take at least a month for a human just to count to a million, but likely longer than that.
When Summit ran thousands of simulations, it identified 77 drug compounds that might be effective in stopping the virus from infecting host cells. Of course, clinical tests need to be performed, which takes time. How much time is unclear. In the United States, Dr. Anthony Fauci, director of the National Institute of Allergy and Infectious Diseases, told reporters it could take 12 to 18 months to develop a vaccine for public use.
You have to go out and test it and ensure that it works,” said Dr. Chuck Gerba, a professor of environmental science at the University of Arizona, referring to a vaccine. He is an expert on the spread of germs. “That takes time, even if you have one developed fairly rapidly within a few months, to really get the data back on its efficacy would probably take a year, even with maybe a benefit of using artificial intelligence.”
Still, the race is on, as biotech companies like Johnson & Johnson, Moderna Therapeutics and others work on a vaccine. While there’s still much testing to be done, Moderna said an emergency vaccine for health care workers could be available as early as this fall. “Any emergency use would be subject to authorization by the appropriate regulatory agencies, based on the emergence of clinical data for mRNA-1273 that would support use of the vaccine prior to licensure,” the company said in a written statement on its website.
Another interesting AI application is in the area of CT scans. With a shortage of PCR tests available, countries are scrambling to figure out what to do. British researchers have even developed a “rapid smart testing device for coronavirus.” According to the Lancaster University website, “Incorporating artificial intelligence, image processing, molecular virology and vast experience in the applied technologies, researchers at Brunel University London, Lancaster University and University of Surrey have developed an innovative device to detect COVID-19 in 30 minutes using an intelligent smartphone application.”
The test, the article says, has already been used for another purpose: checking chickens from the Philippines for viral infections.
For COVID-19, a nasal or throat swab is taken and then put into the device, which costs about $120 each. Within 30 minutes, the researchers say, the device will tell if the sample if positive or negative for COVID-19.
Separately, Abbott Laboratories recently announced a portable testing device called the ID Now Platform that it says can detect the virus in “as little as five minutes.” In a news release, the company said it plans to produce 50,000 tests per day. “With rapid testing on ID Now,” the company said in a written statement, “health care providers can perform molecular point-of-care testing outside the traditional four walls of a hospital in outbreak hotspots.”
Beyond PCR testing, a company called RADLogics, which is based in Boston and Tel Aviv, Israel, is using artificial intelligence through CT scans to help determine if someone has COVID-19.
CT scans are “much faster in detecting COVID-19 symptoms,” wrote Moshe Becker, RADLogics’ CEO and co-founder, in an email. “The average time it takes to do a chest CT is one to two minutes, and our software can analyze the data in less than a second to produce a full analysis including positive/negative determination and quantification of the symptoms if positive.”
The scan, he said, is designed as an automatic detection tool that will help caregivers assess the severity of the spread of COVID-19 in the patient. The software then creates a “Corona Score” based on what percentage of a patient’s lungs are infected with COVID-19. This allows doctors to better assess the severity of the illness, Becker said.
The technology has been used in China, Russia and Italy, he said, and, a study published on the Cornell University website found that, in 157 patients from China and the United States, the test was able to identify COVID-19 and non-COVID-19 patients with an Area Under the Curve (AUC) of 0.996. The best possible AUC is 1.0 in terms of accuracy.
“Given the scale of this pandemic, innovation is more important than ever, and we believe that health care needs to explore all diagnostic and therapeutic options,” Becker added.
SEARCH AND RESEARCH
AI is also playing a role with scientific papers as well. The website Semantic Scholars offers an easy way to track research on COVID-19.
Beyond testing, the need to be able to assess and process information is vital. The U.S. Centers for Disease Control and Prevention now has a “self-checker” guide it calls “Clara” on its website to help people “make decisions about seeking medical care” for COVID-19. AI is also playing a role with scientific papers as well. The website Semantic Scholars offers an easy way to track research on COVID-19.
By using natural language processing technology, the COVID-19 Open Research Dataset can help researchers find information from, at this writing, from thousands of scholarly articles written about the virus. There’s also an “AI-powered Research Feed” that helps scientists “find papers relevant to your interests based on your ratings.”
Additionally, there’s the Delphi research group from Carnegie Mellon University’s Machine Learning department that Roni Rosenfeld, an adjunct professor in the department, runs. The research group uses machine learning data to make predictions about the spread of the flu virus. The group’s past predictions, based on machine learning, have been so accurate that, in October 2019, the U.S. CDC awarded it $3 million in research funding and a five-year designation as an Influenza Forecasting Center of Excellence. “The CDC now routinely includes our forecasting in their messaging to the public and to decision-makers,” Rosenfeld said in an article on the Carnegie Mellon website.
But when the CDC came back to Rosenfeld to ask about tracking the COVID-19 outbreak, he told Vox that he was “very, very reluctant.” The problem with COVID-19 forecasting, he said, is “there is no historical data to go on. So the machine-learning-based approaches are actually the worst here.”
He eventually decided there is a way to offer some predictions—through what he calls the “wisdom of the crowd” through Delphi’s Crowdcast webpage, which even includes a leader board. Basically, volunteers offer their predictions based on common-sense reasoning from the news, from anecdotal knowledge, etc. The people who participate are not experts in making predictions on the spread of a virus, he told Vox, but “what we’ve learned from experience is that any one of them on their own is not very accurate, but their aggregate tends to be quite accurate.”
In terms of testing Italy and South Korea were well ahead of the rest of the world in terms of tests per million people through March 31, according to Our World Data. Testing, of course, allows people peace of mind if they don’t have the virus but also allows people who are carrying the virus but are asymptomatic to know they should quarantine themselves so as not to spread the virus to others.
Everyone, for example, in the small Italian city of Vò, where the first COVID-19 death occurred in Italy, has been tested. Vò is a little over an hour west of Venice. I highly recommend dropping the little guy from Google Street View into Vò and doing some exploring. The University of Padua tested all the people who live there—3,300 in all—for COVID-19. Twice. and learned that “at least six” people tested positive who were asymptomatic. In Vò, “there were the first two cases,” Luca Zaia, the governor of Italy’s Veneto region, told the Guardian. “We tested everyone, even if the experts told us this was a mistake: 3,000 tests. We found 66 positives, who we isolated for 14 days, and after that six of them were still positive. And that is how we ended it.’’
It has been well-documented that the United States is already behind in COVID-19 testing and now faces another challenge: a shortage of supplies to perform the tests. It generally takes three or four days to get test results, but the U.S. Food and Drug Administration did recently approve a new test that can detect the virus in about 45 minutes to go along with the previously mentioned Abbott five-minute test. And with testing, speed is of the essence.
And that need for speed starts with our very DNA. “Whenever I’m teaching I always give an example that when I did my Ph.D., and this was 2000 to 2004,” UIC’s Green said, “I sequenced 150 genes as part of my Ph.D. thesis.
Nowadays, we have sequencing runs that generate 150 million sequences, a billion sequences, 10 billion sequences. So, the magnitude of the data output are so incredible compared to what was available even 10 years ago, even five years ago.
“Advancements in DNA research include novel sequencing chemistries and methodologies, miniaturization of sequencing instrumentation allowing millions of more reactions to be conducted concurrently, and greater computational capacity to acquire, process, store and distribute sequence data,” Green said. “In fact, the output of sequence data has become so great that computational analysis is often the bottleneck, rather than data acquisition.”
In terms of developing anti-viral medications, which could be used to treat COVID-19 patients, one interesting area is virtual reality.
Dr. Adrian Mulholland is a chemistry professor at the university of Bristol’s Center for Computational Chemistry, in England. He and his team are using VR to help create medicines. They’ve researched influenza proteins among others to learn how drug molecules bind to them. And now, “We want to apply this to COVID-19 targets,” he told Spark.
It could produce something like a Tamiflu drug that could be used prophylactically for health care workers to help prevent them from getting COVID-19 or it could be used “to slow progression of the disease, to mitigate the symptoms and limit infectivity.”
”When researchers put on the VR headset, Mulholland said, “you really step into the molecular world, you step inside the protein that’s the target for the drug design, and you see this in front of you.”
With the VR controller, you can manipulate the molecules and see what works to combine a drug to the virus to help prevent the functionality of the virus.
Besides trying to create new medications, the researchers are also testing how drugs that are currently available could perform in treating COVID-19.
“You can scan through maybe thousands of potential drugs at a time,” he said, referring to research into using available medications that treat other diseases. “But a quick and dirty approach like that will generate a lot of false positives. So we want to use the VR tool to help eliminate those false positives to identify the really promising candidates.”
Mulholland also wants to provide the technology to more researchers around the world through an open-source framework, which opens up a new dimension of research possibility.
Speaking of open source, Spark caught up with Colin Keogh, a mechanical engineer who co-founded the Open Ventilator Project in Ireland. The project is using crowdsourcing and group think to create a prototype that health care providers could possibly use to print emergency ventilator systems in a pinch. As more people become ill with COVID-19, ventilators will play a huge role in the recovery process. The project is designed to help fill the void should a ventilator shortage continue or get even worse.
“We’re not trying to replace standard hospital ventilators,” Keogh said. “We’re not trying to develop new forms of medical ventilators. We’re just trying to provide a solution that can be used in an emergency when it’s either a choice between this or nothing.” The world, he added, hasn’t faced a challenge like this, probably since World War II. He believes that regulations could be relaxed during a global emergency, so he wants to be prepared with a solution if that should happen.
“We don’t want to over promise anything,” he said. “We want to test our solutions first and make sure it is a valuable solution. So we’ve been working an awful lot with doctors and medical practitioners, getting their advice and input on the concepts as they currently stand. And then, as the program develops and as we iterate these concepts and designs, we will probably have some interactions” with governments to see the potential viability of open-source ventilators.
Swimming right along, similarly, Italian engineers recently used 3-D printing to turn scuba gear into ventilator equipment on a limited scale. Of course, patients would have to sign a waiver to use the uncertified device, according to an article on Futurism.com. Dyson, the vacuum and hand dryer company, has even gotten into the ventilator game. James Dyson, the founder, was able to design a new ventilator in just 10 days and plans to make 15,000 of them.
The antithesis of information is misinformation
While social media is a powerful tool to get the message out, it’s also just as powerful as an instrument of misinformation. Seven of the biggest social media platforms, including Facebook, Twitter and YouTube, issued a joint statement on March 16 that read: “We are working closely together on COVID-19 response efforts. We’re helping millions of people stay connected while also jointly combating fraud and misinformation about the virus, elevating authoritative content on our platforms and sharing critical updates in coordination with government health care agencies around the world. We invite other companies to join us as we work to keep our communities healthy and safe.”
Facebook, for one, has started a Coronavirus (COVID-19) Information Center for the latest updates, and says it will connect its users to trusted sources and work to weed out misinformation among other steps, according to a post attributed to Kang-Xing Jin, Facebook’s head of health. And, as TechCrunch reports, the popular instant messaging platform WhatsApp, which is owned by Facebook, has been testing a feature that allows users to get more context and information about the messages they receive to help people fi lter misinformation.
Anything that tamps down on untruths, like the list that Dr. Faheem Younus, a physician who is chief of infectious diseases at the University of Maryland Medical System, debunked on Twitter, gets a thumb’s up from Spark. No, hot water and onions cannot prevent someone from getting COVID-19. Got it?
In general, Keogh from the Open Ventilator Project, believes a time of crisis is also a great time for innovations.
“So, again, to reference the Second World War,” Keogh said, “the amount of technology that was created for that or the Cold War or the space race, whenever there is this larger collective drive to either solve a problem or do something new, it does seem to stimulate people’s innovation capacity.”
Exactly right, and, if you’ll excuse me, the dog is barking. Stay safe.