Virus World

Machine-learning system trained on millions of human audio clips shows promise for detecting COVID-19 and tuberculosis.  A team led by Google scientists has developed a machine-learning tool that can detect and monitor health conditions by evaluating noises such as coughing and breathing. The artificial intelligence (AI) system, trained on millions of audio clips of human sounds, might one day be used by physicians to diagnose diseases including COVID-19 and tuberculosis and to assess how well a person’s lungs are functioning. This is not the first time a research group has explored using sound as a biomarker for disease. The concept gained traction during the COVID-19 pandemic, when scientists discovered that it was possible to detect the respiratory disease through a person’s cough.

What’s new about the Google system — called Health Acoustic Representations (HeAR) — is the massive data set that it was trained on, and the fact that it can be fine-tuned to perform multiple tasks. The researchers, who reported the tool earlier this month in a preprint1 that has not yet been peer reviewed, say it’s too early to tell whether HeAR will become a commercial product. For now, the plan is to give interested researchers access to the model so that they can use it in their own investigations. “Our goal as part of Google Research is to spur innovation in this nascent field,” says Sujay Kakarmath, a product manager at Google in New York City who worked on the project....

Preprint available here https://arxiv.org/pdf/2403.02522.pdf 

Researchers at the top-ranked hospital in Boston compared search interest in loss of taste and appetite, and diarrhea with the reported incidence of Covid-19 in 15 U.S. states from January 20 to April 20. Using Alphabet Inc.s Google Trends online tool, they found the volume of searches correlated most strongly with cases in New York, New Jersey, California, Massachusetts and Illinois - states with high disease burden - three to four weeks later. Internet searches on gastrointestinal symptoms predicted a rise in Covid-19 cases weeks later, researchers at Massachusetts General Hospital found, demonstrating a novel early warning system for hot spots of the pandemic disease. 

The research, published in the journal Clinical Gastroenterology and Hepatology, showed that the same approach used to monitor pandemic influenza trends more than a decade ago could be deployed for the coronavirus, the hospital said in a report this month. Patients with Covid-19 often report gastrointestinal symptoms, such as abdominal pain and diarrhea, sparking interest in conducting the study. “Our data underscore the importance of GI symptoms as a potential harbinger of Covid-19 infection and suggests that Google Trends may be a valuable tool for prediction of pandemics with GI manifestations,” Kyle Staller, a gastroenterologist and the director of Mass General’s gastrointestinal motility laboratory, and colleagues wrote in the study. Scientists are also testing for traces of the coronavirus in wastewater to identify places where Covid-19 is spreading.

Google has added several Covid-19-related features and alerts to its Maps app for iOS and Android. Traveling during the pandemic can be tough. That's why the latest version of Google Maps for Android and iOS has several features that might help travelers get around safely. Now, when you look up public transit directions, Maps will show COVID-19-related alerts. Google points out that data will only be available where it can get info from the local transit agencies, and only on trips that are likely to be affected by COVID-19 restrictions. These alerts are rolling out in Argentina, Australia, Belgium, Brazil, Colombia, France, India, Mexico, Netherlands, Spain, Thailand, United Kingdom, and the U.S., and Google says they're coming to more countries "soon." 

Maps will also display alerts when you navigate to medical facilities or COVID-19 testing centers, warning you to "verify eligibility and facility guidelines to avoid being turned away or causing additional strain on the local healthcare system." Alerts for medical facilities are rolling this week in Indonesia, Israel, the Philippines, South Korea, and the U.S. Alerts for testing centers will only be available in the U.S. Again, Google says it will only show these alerts when it can get "authoritative data from local, state and federal governments or from their websites." Google says the app will also tell you how crowded buses, subways, and transit stations will be. Plus, you'll be able to see when a transit station has historically been less busy so you can plan ahead. Google says these features are powered by aggregated and anonymized data from users who opt in to Google Location History. They'll be rolling out in the "next several weeks."

Finally, the app will show you driving alerts, notifying you about COVID-19 checkpoints and restrictions along your route — for example, when crossing national borders. This feature will initially only be available in Canada, Mexico and the U.S. 

Google Maps will soon show how prevalent coronavirus is in geographic areas with a new color-coded update. Beginning this week, the Maps app will display seven-day averages of new Covid-19 cases per 100,000 people. The chosen areas will show if cases are increasing or decreasing and be shaded with one of six colors to signify how many new cases were reported. 

In a blog post. Google said the tool shows "critical information about Covid-19 cases in an area so you can make more informed decisions about where to go and what to do." Google is pulling data from three sources: Johns Hopkins University, the New York Times, and Wikipedia. They receive their data from the World Health Organization and other public or government health organizations. Information is available for all 220 countries that Google Maps works in. The feature will soon be available on iOS and Android versions of Google Maps. To use it, open the app, press the layers button (two squares on top of each other) on the top right and press "Covid-19 Info" button. That will reveal the color-coded app showing case counts. This week's update is just the latest for the ever-evolving app. Last month, it rolled out a colorful and more detailed redesign that aims to help people more easily distinguish different types of terrain. Google Maps also placed a greater emphasis on discovery and recommendations, along with a new logo, in a February update to celebrate its 15th birthday.

Verily is opening its own testing lab. The lab is in-house at the Google sibling company's South San Francisco headquarters. Verily, the life sciences arm of Google's parent company Alphabet, on Tuesday said it has set up its own coronavirus testing lab aimed at getting people faster results. The lab, located on the company's campus in South San Francisco, California, was built to run "several thousand" tests per day, Verily said. The announcement comes as COVID-19 cases have spiked, causing labs to return results at a slower clip. Facilities that used to take two to three days are now taking a week or longer.  

"When the pandemic hit, it became clear that we needed to rapidly establish a lab and to receive California state licensure and CLIA certification, which we have done," said Verily Head of Pathology Deb Hanks, referring to the state's standards for clinical labs. "We've established this lab to provide a focused specialty service with rapid turnaround time."  One way Verily says it's increasing testing capacity is by looking into "pooled" testing, or combining respiratory samples from several people and conducting one lab test on the sample set. The Verily facility will focus primarily on customers of the company's Healthy at Work program, aimed at helping businesses and schools get people back to offices or campus. 

Verily's efforts were the source of drama and confusion earlier during the pandemic, after President Donald Trump announced the federal government was working with Google on a coronavirus testing website. The tool, which was announced unexpectedly by the president, turned out to be a website that allows people to take a screener survey to see if they should go to testing stations for COVID-19. The screening website has drawn privacy scrutiny from lawmakers. In order to take the online screener, the site requires people to sign in using a Google account. Democratic senators have pressed Verily about the issue, but in April, Verily indicated it would keep the requirement for security and authentication reasons.

Alphabet's life sciences arm Verily unveiled a coronavirus antibody research study to help determine the accuracy of serology tests— a research initiative aimed at better understanding coronavirus antibody testing. Verily is adapting its existing Project Baseline study to support the antibody research endeavor, the first initiative of which will be to roll out out serology testing to patients who have already taken a nasal swab test from Verily's testing program. For context, Verily erected four coronavirus testing sites across California early last month.

Verily's study could be instrumental in establishing the accuracy of coronavirus antibody tests and will add to its myriad coronavirus-specific healthcare initiatives. There are wide variations in the accuracy of different antibody tests — and Verily's research initiative could help the medical community better understand coronavirus immunity as many states prepare to reopen. For context, a number of vendors have released their own antibody tests amid the pandemic, but only a handful have gained FDA approval for use.

Researchers at UCSF and UC Berkeley have noted that dozens of tests performed reasonably well in determining antibodies within the first two weeks of infection — only to produce a greater number of false positives that exceed the number of infected people in some regions. As some states start to resume normal operations, they may rely on using positive antibody test results as "immunity passports" enabling consumers to return to work — but researchers are still uncertain whether a positive result guarantees immunity. And the discrepancies in test results add another layer of ambiguity, as serology tests have not yet been proven to be a reliable marker in determining whether an individual has coronavirus antibodies...

Verily's announcement (May 18, 2020): 

https://blog.verily.com/2020/05/new-baseline-covid-19-research-project.html