AI is dreaming of drugs no one has ever seen. Now we have to see if they work.
Today, on average, it takes more than 10 years and billions of dollars to develop a new drug. The vision is to use AI to discover drugs faster and cheaper. By predicting how potential drugs might work in the body and weeding out worn-out compounds before they leave the computer, machine learning models can cut down on the need to perform hard work in the laboratory.
“There are still too many diseases we can’t treat or can only treat with a three-mile list of side effects,” said Adityo Prakash, CEO of California-based pharmaceutical company Verseon. .”
Now, new laboratories are being built around the world. Last year Exscientia opened a new research center in Vienna; In February, Insilico Medicine, a drug discovery company based in Hong Kong, opened a new large laboratory in Abu Dhabi. All told, about two dozen (and counting) AI-assisted drugs are currently or are in the process of clinical trials.
“If someone told you they could perfectly predict which drug molecule might pass through the gut… they might as well have land to sell you on Mars.”
Adityo Prakash, CEO of Verseon
Sean McClain, founder and CEO of Absci, a Vancouver-based company, explains: “We are seeing this increase in activity and investment because of the automation boost. Chemistry in the pharmaceutical industry has begun to generate enough chemical and biological data to train good machine learning models. Washington, uses AI to search through billions of potential drug designs. “Now is the time,” McClain said. “We will see massive transformation in this industry over the next five years.”
However, it is still early to discover AI drugs. “There are a lot of AI companies that make claims that they can’t back it up,” says Prakash: “If someone told you they could perfectly predict which drug molecule could pass through the intestines or not be broken down by the liver. cancel, stuff like that, they might have land to sell you on Mars too.
And the technology is not a panacea: laboratory cell and tissue experiments and human trials—the slowest and most expensive parts of the development process—cannot be completely ruled out. whole. “It saves us a lot of time. It took a lot of steps that we used to do manually, says Luisa Salter-Cid, chief scientific officer of Pioneering Medicines, part of the Flagship Pioneering startup incubator in Cambridge, Massachusetts. “But final validation needs to be done in the lab.” However, AI has changed the way drugs are made. It may be a few years before the first drugs designed with the help of AI hit the market, but the technology is set to shake up the pharmaceutical industry, from the early stages of drug design. up to the final approval process.
The basic steps involved in developing a new drug from scratch haven’t changed much. First, choose a target in the body with which the drug will interact, such as a protein; then design a molecule that will do something with that goal in mind, such as change the way it works or turn it off. Next, create that molecule in the lab and check if it really does what it’s designed to do (and nothing else); and finally, test it in humans to see if it’s safe and effective.
For decades, chemists have screened potential drugs by placing samples of the desired target in many small compartments in the laboratory, adding different molecules, and monitoring the reaction. They then repeated the process many times, tweaking the structures of the candidate drug molecules — swapping one atom for the other — etc. Automation sped things up, but the trial and error process. Core mistakes are inevitable.