AI imagines drugs that 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 make drug discovery faster and cheaper. By predicting how potential drugs might behave in the body and eliminating dead-end compounds before they leave the computer, machine learning models can reduce the need for painstaking lab work.

And there’s always a need for new drugs, says Adityo Prakash, CEO of California-based pharmaceutical company Verseon: “There are still too many diseases that we can’t treat or can only treat with drug lists. side effects three miles long. .”

Today, new laboratories are being built all over the world. Last year, Exscientia opened a new research center in Vienna; In February, Insilico Medicine, a Hong Kong-based drug discovery company, opened a large new laboratory in Abu Dhabi. In total, about two dozen drugs (and counting) that have been developed with the help of AI are currently in clinical trials or are entering clinical trials.

“If someone tells you they can perfectly predict which drug molecule can get through the gut…they probably have some land for sale on Mars too.”

Adityo Prakash, CEO of Verseon

We are seeing this increase in activity and investment because increasing automation in the pharmaceutical industry has begun to produce enough chemical and biological data to train good machine learning models, says Sean McClain, founder and CEO of ‘Absci, a Vancouver, Washington-based company that uses AI to search billions of potential drug designs. “Now is the time,” McClain says. “We’re going to see a huge transformation in this industry over the next five years.”

Yet it is still in its infancy for AI drug discovery. There are a lot of AI companies that claim they can’t back it up, says Prakash: “If someone tells you they can perfectly predict which drug molecule can go through the gut or not be broken by liver, things like that, they probably have land to sell you on Mars too.

And the technology is no panacea: experiments on cells and tissues in the laboratory and testing on humans – the slowest and most expensive parts of the development process – cannot be eliminated entirely. “It saves us a lot of time. It already does a lot of the steps we used to do by hand,” says Luisa Salter-Cid, chief scientific officer of Pioneering Medicines, part of startup incubator Flagship Pioneering in Cambridge, Massachusetts. “But the ultimate validation has to be done in the lab.” Yet AI is already changing the way drugs are made. It may be a few more years before the first drugs designed with the help of AI hit the market, but the technology is poised to disrupt the pharmaceutical industry, from the early stages of drug design drugs 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 to that target, like change how it works or shut it down. Then, make that molecule in a lab and verify that it actually does what it was designed to do (and nothing else); and finally, test it in humans to see if it is both safe and effective.

For decades, chemists have screened for drug candidates by placing samples of the desired target into many small compartments in a lab, adding different molecules, and watching for a reaction. Then they repeat this process several times, adjusting the structure of the candidate drug molecules – replacing this atom with this one – and so on. Automation has sped things up, but the basic process of trial and error is unavoidable.