Getting ahead of the next pandemic with AI and machine-learning fueled drug discovery

The Rapidly Emerging Antiviral Drug Development Initiative (READDI) uses advanced technologies to prepare therapeutic drugs for future viral outbreaks

Coronavirus. Alphavirus. Flavivirus. These and other virus families are keeping virologists up at night for their pandemic potential and lack of vaccines or treatments.

READDI, a public-private biotechnology company, uses historical and medical data, analytics and artificial intelligence to identify which virus families are most likely to cause major outbreaks of new viruses and develop new antiviral drugs before they are needed.

A mission to be pandemic-ready, not reactive

Tackling a global problem requires a global effort. READDI brings together the world’s best scientific and business minds to prevent infectious disease threats from turning into a pandemic like COVID-19.

READDI began as a collaboration between the University of North Carolina at Chapel Hill, Eshelman Institute for Innovation at UNC-Chapel Hill and the Structural Genomics Consortium. The effort expanded with collaborations among academic research institutions, industry, government and philanthropic organizations. Together, they rallied behind READDI’s mission to develop effective antiviral drugs to stop current and future pandemics.

In early 2020, when virologists learned COVID-19 was spreading asymptomatically, they knew the world wasn’t prepared for COVID-19 any more than it had been for the SARS, Ebola, and Zika viruses.

“We knew we weren’t going to have the drugs or vaccines we needed,” says Dr. Nat Moorman, READDI Co-founder and Associate Professor in the UNC School of Medicine’s Department of Microbiology and Immunology. “We were very concerned this was going to turn into a global pandemic. That’s when we decided to form READDI and start making drugs in advance so we wouldn’t have to deal with this again.”

Using advanced technologies to transform drug development

READDI has partnered with SAS since 2021 to apply the most advanced technologies to transform the drug development process. The goal is developing broad-spectrum, small molecule antiviral drugs before they’re needed, instead of starting from scratch when a new virus emerges.

READDI’s antiviral expertise combined with SAS technology not only prepares for future viral outbreaks but also brings more power to develop better treatments for existing viruses such as COVID-19.

“We’ve already seen the power of bringing together READDI’s collected data with SAS Viya and our data scientists,” says Rachel Hardin, Head of Market and Business Development for US Life Sciences at SAS. “We know there’s much more to uncover related to novel target discovery, not only with COVID-19 but with other viral diseases. We see this as a great future opportunity to test the limits of Viya and see how we can continue to support this AI-driven drug discovery platform with READDI’s help.”

Two broad-spectrum antivirals per high-risk family

When a new virus infects humans, our immune systems are powerless to stop it from replicating. Viruses in the same family share inherited methods for hijacking cells to essentially become virus factories.

READDI scientists identify those common factors within virus families and exploit them as a viral Achilles heel. “Most antiviral drugs are designed to work against one virus,” Moorman says. "The problem when we think about pandemics and epidemics is we don’t know what virus comes next. We’re aiming to make a single drug that works against multiple related viruses to provide treatments for current viruses and protection from future threats.”

READDI set a goal to develop two broad-spectrum, small molecule antiviral drugs – pills that can be taken with a drink of water — per high-risk virus family to prevent severe illness, hospitalization and death.

Those pills would serve as a frontline defense against emerging virus outbreaks, in either the absence of or complementary to vaccines.

AI and machine learning foster collaboration

READDI initially teamed with data scientists in the SAS Analytics Center of Excellence for help managing large volumes of data. “By bringing the data into our Viya environment, we could begin to discern trends using key analytic tools within the platform,” Hardin says.

The partnership evolved to include cloud-based AI and machine learning, which help READDI teams discover drugs faster while fostering secure collaboration between research partners. READDI and SAS use analytic platforms and AI models to rapidly cycle through and find the best small molecule compounds.

SAS research and development teams applied advanced machine learning techniques to integrate multiple biological data sets READDI collected from severely ill COVID-19 patients. UNC researchers combined that with analysis of Electronic Health Record data and applied AI modeling to identify targets for new antiviral drugs.

“It’s been exciting to see new things we never would have thought to look at become critical in some cases to our understanding of disease,” Moorman says. “It’s been working so well because of the SAS team’s willingness to engage, understand what we’re trying to do and then translate that into the AI models.”

A new world of precision therapeutics

READDI’s work with SAS and COVID-19 patients showed that people vary in their host response to the same virus. Patients may need treatment tailored to the way they particularly develop disease to be fully effective. What began as developing broadly active antivirals could become more targeted as READDI’s understanding of each patient’s response grows.

“The work we’ve been doing with SAS is going to change the way we think about viral diseases,” Moorman says. “Understanding how the host response drives disease is opening a new world of precision therapeutics in antivirals. We’re creating a fundamental paradigm shift in how we approach treating viral diseases with precise medicines to give to the right patient at the right time.”