The trend toward digital transformation of production, which has been actively developing in recent years, has not bypassed the pharmaceutical industry. Since 2005, the number of cases of artificial intelligence implementation in medical processes has grown almost 62 times. Kamila Zarubina, Director for Acceleration and Partner of the Biomedical Cluster of the Skolkovo Foundation, spoke about how such technologies can improve human health and speed up drug development.
The Art Of Pharmaceutical Intelligence
Artificial intelligence (AI), used in pharmaceuticals and other bioindustries, is a highly specialized machine intelligence. It is designed to solve specific problems using automated algorithms. Its goal is to find hidden patterns and gather information from huge amounts of data in ways that are inaccessible to humans. Applications range from industrial process automation and clinical applications to drug discovery.
The development of a drug has several stages:
- Search for a molecule or drug candidate;
- Search for a target;
- preclinical and clinical studies.
AI algorithms can be used at every stage of development. For example, to analyze a molecule to make a preliminary model of its effect on the target. Next, check its mechanism of action on certain human organs. After that, preclinical and clinical studies are modeled.
Everyone knows that drug development is very long and expensive. And the entire pharmaceutical industry is aimed at reducing the risk to humans to the minimum possible. Classical methods for testing a drug for safety take a long time. And mathematical modeling algorithms can reduce the development time without compromising future drug safety.
Faster, Better, More Accurate
Thus, we can immediately talk about two positive effects artificial intelligence has on the pharmaceutical industry. Firstly, it reduces development time, and secondly, it allows you to optimize costs and lower the final cost of the drug. Several residents are working on AI services to simulate the drug development process.
For example, SaaS platforms help organize clinical and observational studies. Today, this service brings together more than four thousand doctors in 125 cities across the country. With its help, scientists can share best practices, prepare studies faster, find participants, and optimize the budget. At the same time, special software reliably protects patient data.
Another striking example from among the participants in the Skolkovo project is Insiliko. The world-famous company develops software that integrates omics data and deep learning to assess the toxicity, pharmacokinetic properties, and effects of drugs on the body. Another resident, OncoyNite, is developing test systems for early detection and monitoring melanoma treatment effectiveness. Now she is also moving to other nosologies.
Doctors also often face the problem when new diseases or subsequent strains of already known infections become increasingly resistant to antibiotics. Living organisms adapt to the environment and survive, despite adverse factors. The same thing happens with diseases – antibiotics eventually cease to work on bacteria as they adapt. As a result, there is no cure. I want artificial intelligence to help us with this.
However, this is just a promising direction so far – it is still difficult to name any specific products ready for implementation in real clinical practice. But specially trained neural networks could identify patients with a high chance of developing antibiotic resistance and advise an analog of the drug, which in this case would be better suited. Plus, AI algorithms can accelerate the development of next-generation antibiotics – but this is already a medium-term perspective with a horizon of about ten years. And in general, this is closer to predictive medicine.