From ‘The Sims’ to ‘In Silico Clinical Trials.’ How AI Ensured the Success of the COVID Vaccine.
On Tuesday, February 26, 2019, a little after midnight, thousands of gamers downloaded “StrangerVille,” the most recent add-on for the life simulation franchise “The Sims.” Reviews were mixed: Some critics thought Electronic Arts had, finally, milked the cash cow dry. One reviewer called the game pack “empty.”
“StrangerVille” built another elaborately implausible story arc: Something about scientists and conspiracy theories and mutant plants. In the game, players test vaccines at the “Chemical Analyzer” in the basement of a secret military lab, then head to town and vaccinate infected Sims. “The vaccine probably won’t work at first, but you can learn a bit from it,” noted one reviewer.
The vaccination role-play, nor the game itself, made little impact. Most gamers forgot about it when Activision released “Call of Duty: Modern Warfare” a few months later.
Then the pandemic hit.
In the months that followed, scientists raced to find treatments for COVID-19, which brought the greatest global health crisis for generations. But before vaccine trials began, biopharma companies had to test the efficacies of potentially life-threatening drugs in virtual experiments. These tests involved no patients, just pixels. Using advanced computer modeling, scientists would identify the toxicity of chemicals without injecting, and potentially killing, humans. (Or animals.)
And suddenly, “StrangerVille” didn’t seem so implausible after all.
Scientists did what gamers did the year before: Infecting, then vaccinating, virtual patients in a simulated environment — something called in silico clinical trials. The only difference? Gamers progressed to the next level. Scientists saved the world.
You know the rest. As of March 2021, millions of people worldwide have received a COVID vaccination. But few people know that in silico screening played such a critical part in the pandemic.
This is that story.
What are In Silico Clinical Trials?
An in silico screening clinical trial (or in silico testing or in silico toxicology) is a tough concept to explain. The Sims example, where scientists, like gamers, manipulate outcomes virtually, is probably a good place to start. Just know: There’s an incredible amount of AI involved.
In a COVID context, biopharma plowed resources into in silico. The world demanded a vaccine, but a rushed rollout could damage the global population more than the virus itself. Thousands of people signed up for trials, but scientists couldn’t put injections in volunteers’ arms unless treatments were safe.
For in silico to work, scientists need data. And lots of it. Then they transfer this data to a computational screening system, which uses in silico methods like simulation and visualization to create super-smart computer models that test in vivo and in vitro data in virtual environments.
In silico provides scientists with unparalleled insights such as how chemical properties impact drug efficacy or why patients might absorb drug composites differently. Scientists can then eliminate undesirable compounds and test drugs in the real world.
A (Very) Brief History of In Silico
In silico came out of necessity. Western governments banned animal testing as early as the ’90s, and while the United States still has no federal laws prohibiting this practice, biopharma has known the ethical implications for decades. So with no animals to test, bioscientists turned to computer scientists, who developed highly complex simulation systems that powered drug research and development.
Advancements in computer science and the legislative and ethical response to animal testing formed a perfect synergy, allowing scientists to develop, then eventually deploy, drugs in the ’90s and ’00s. But in the ’10s, AI changed the game. Now scientists could create simulations based on a mammoth range of data sets — DNA sequences, electronic health records, socio-economic records, you name it. They fed all this data into computer systems like Professor Frink did in “The Simpsons,” generating accurate models that prefaced the most significant medical trials of the decade.
COVID-19 and In Silico
On April 1, 2020, the world had pretty much ground to a halt. Stores had rolled down the blinds and closed the doors. Airplanes lay dormant on airstrips. Federal, state, and local governments issued stay-at-home orders. “Social distancing,” “furlough,” and “unprecedented” had entered the collective lexicon. Twenty percent of the American population was working from home. (This number would increase the following month.)
There was more demand for a vaccine than at any point in human civilization.
So biopharma had two options:
- Rush head-first into human clinical trials and hope for the best, potentially risking lives.
- Or invest more time and resources into in silico, which evaluated the safety of chemicals before treatments reached the human clinical trial stage.
Biopharma chose the latter.
Other testing and screening methods were paramount to COVID vaccine development. But in silico heavily relies on AI, which makes it such a fascinating subject for technophiles. It’s all about math and models and manipulating data. There’s a certain science-fiction feel to in silico, where complicated computer algorithms have the power to end a pandemic. But in silico is very much science fact.
There are many examples of researchers using this tech in 2020. Some scientists, for example, developed a comprehensive model that incorporated elements of the adaptive and innate immune systems, rates of viral replication, the system responsible for cellular entry inflammatory, cytokines, and something called the “coagulation cascade.” (It relates to blood clots.) The model discovered clinical outcomes and treatment responses to understand vaccine patients. Other scientists developed an analytical platform that examined serological and vaccine targets.
Before You Go
It’s hard to estimate the impact of in silico screening on COVID vaccine development. It’s just one stage of the evolution of the vaccines that nurses are injecting into patients’ arms right now. But expect computer simulations to play a more significant role in the future of medical science, where biopharma companies test drugs in virtual environments before administering them to the public. Just like in “StrangerVille.”