Computational Crystal Ball Predicts Which COVID Variants Are More Dangerous

Viruses are nature’s shapeshifters—their genomes are in a constant state of flux as they divide in their hosts. Over time, they gradually acquire a library of errors that eventually change the way the virus behaves.

This evolutionary phenomenon has led to the emergence of SARS-CoV-2 variants such as Delta; a COVID variant thought to be more than twice as contagious as previous iterations of the coronavirus.

Given that the virus always seems to be one step ahead of efforts to control the pandemic, researchers have been seeking ways of predicting exactly how SARS-CoV-2 is changing over time and what that means from an epidemiological standpoint. Such platforms would buy us time, signaling the potential emergence of a potentially more dangerous variant before it sweeps through and devastates communities.

To this end, scientists at Penn State have made a giant leap forward, creating the first computational framework for predicting genetic changes in the coronavirus’ spike protein. The virus uses this protein as a molecular key to gain entry into human cells via their ACE2 receptors.

The vast majority of circulating variants possess mutations in the spike protein’s receptor binding domain (RBD). These amino acid changes boost the infectivity of the virus, allowing SARS-CoV-2 to enter and infect human cells faster and more easily, making the virus more transmissible. These accumulated changes in the RBD may also help variants evade the immune systems of vaccinated individuals or those who have recovered from COVID-19.

In their study, the researchers developed a two-step computational framework that models changes occurring in the RBD and links these mutations to the ability of the virus to bind to human ACE2 receptors.

“The binding strength between RBD and ACE2 directly affects infection dynamics and potentially disease progression,” explained the lead investigator of the study, Suresh Kuchipudi. 

“The ability to reliably predict the effects of virus amino-acid changes in the ability of its RBD to interact more strongly with the ACE2 receptor can help in assessing public health implications and the potential for spillover and adaptation into humans and other animals.”

Kuchipudi and colleagues leveraged a technique known as Molecular Mechanics-Generalized Born Surface Area (MM-GBSA) analysis, which measures how strongly the virus RBD binds to ACE2. Powered by binding affinity data of existing variants, their new framework uses a deep-learning algorithm to help predict the effects of amino acid changes in the RBD.

The team found that they could predict with over 80 percent accuracy whether or not RBD amino acid mutations improved or worsened the virus’ ability to latch on to ACE-2.

This new framework would empower public health officials with a tool to assess the infectivity of currently circulating and emerging SARS-CoV-2 variants in terms of their infectivity. If a new strain is predicted to have a greater binding affinity to ACE2, countermeasures to protect the public can rapidly be deployed to save lives.