(ORDO NEWS) — In early 2023, it became known that the first drug found by a computer had passed the first phase of clinical trials.
Although its effectiveness is still pending confirmation, many experts hope that a revolution is coming in the field of pharmaceuticals.
Recently, the attention of the whole world has been focused on chatbots with artificial intelligence (AI) – for example, ChatGPT from OpenAI.
But there seems to be another AI revolution going on right now. It’s about using machine learning to find new drugs. Recently, there have been real prospects in this area.
In January 2023, biotechnology company Insilico Medicine announced that its AI-discovered drug had passed Phase I clinical trials.
Similar developments are being carried out by several companies around the world, but this is the first time that “pills from the car” have shown such a result.
Development stalls
Creating a new drug is an extremely long and costly process. Experts need not only to find a chemical compound that targets a specific target in the body, but also to test it.
At any stage of the test, a problem may arise. For example, a drug works well on a target in a test tube, but not in a human body. Or causing side effects. Or the effect is expressed only in animals.
The complexity of human biology makes drug development astronomically expensive.
Back in the 1980s, scientists noticed that the cost of pharmaceutical research and development was doubling every ten years or so.
This effect was later called Eroom’s law. This was a reference to the famous “Moore’s law”, which states that the computing power of computers is growing exponentially.
Only in the case of drugs, it was the other way around (hence Eroom – a mirror image of Moore). The further, the more difficult and expensive it was to find new drugs.
“The number of small molecules is 10 to the 60th power, but only a tiny fraction of this astronomically large chemical space has been explored,” explains Harvard Medical School bioinformatician Marinka Zitnik.
“Developing a drug from scratch that is both safe and effective is incredibly difficult. On average, it takes 11 to 16 years and $1 billion to $2 billion.”
Machine algorithms can take on some of the work: determine which molecules are most likely to be safe and effective in treating humans.
Due to their ability to find patterns (features common to many objects), algorithms “sift” huge amounts of biochemical data.
They can rely on genome sequencing data, assays, and data on the effects of other molecules.
Ultimately, according to the developers, this will allow, if not transferring all the work of finding drugs to machine intelligence, then at least reduce the cost and reduce the time.
The road to breakthrough is just beginning
The INS018_055 molecule found by the neural network should selectively block the work of enzymes from the tyrosine kinase family.
According to one version, they are responsible for the development of idiopathic pulmonary fibrosis.
Scientists synthesized this substance and successfully tested its effectiveness in experiments on mice and other model animals.
The first phase of clinical trials is already the transition to human testing. The drug has already received about 80 volunteers from New Zealand – and it was well tolerated.
But for now, this is just a security check. Phases II and III test effectiveness, first with a small sample of a few hundred people and then with a larger sample of thousands of participants.
They will start in the coming months.
Insilico is not the only startup of its kind. For example, the British company Exscientia hit the headlines in 2021 by announcing the start of the first phase of clinical trials of a drug developed using AI for cancer immunotherapy.
And Utah-based Recursion Pharmaceuticals is using artificial intelligence to find new uses for drugs owned by other companies.
Judging by other areas where AI is actively developing, we can expect an explosive growth in such research in the coming years. However, there may be new difficulties that we do not yet know about.
—
Online:
Contact us: [email protected]
Our Standards, Terms of Use: Standard Terms And Conditions.