AI finds anticancer drug candidates with quantum computing

Innovative technologies such as artificial intelligence (AI), machine learning, and quantum computing can accelerate the design and discovery of new drugs to combat cancer. a New Peer Review Research Published in Natural biotechnology, Leading by the University of Toronto and Insilico Medicine, AI demonstrates how quantum computing can create anti-cancer molecules of proteins that were previously unable to target pharmacologically.

“In my opinion, quantum computing is in the same state as chemistry generation AI from 2015 to 2016 when my team decided to go all-in,” says Alex, founder and CEO of Insilico Medicine. Zhavoronkov said. “We know it’s early. We know we can’t make a profit right away, but we want to push boundary For ourselves and the industry as a whole, we focus on what we know best: chemistry and biology. ”

The global problem of cancer

Cancer is the leading cause of death worldwide and is second only to cardiovascular disease. Our World of Data. Globally, an estimated 35 million cancer cases are projected by 2050. Global Cancer Facts and Figure 5^th Edition By the American Cancer Society (ACS), and about one in five Americans develop cancer in their lifetime.

Cancer is a widespread disease of disease in which healthy cells become cancer due to the growth of uncontrolled cells that can spread. According to the National Cancer Institute (NCI), cancer is gene It manages our cells. Genetic mutations are changes in DNA sequences. Oncogenes are mutant genes that have the potential to cause cancer.

It is important to note that despite cancer being a genetic disease, it is not necessarily equivalent to being a familial or hereditary disease. In fact, according to NCI, the majority of cancers (90% to 95%) are non-headitarian (also known as spontaneous) that is caused by mutations and that lifestyle and environment affect the risk of cancer. For example, cigarettes smokingimmunosuppressants used for tobacco use, certain infectious diseases, radiation, and post-organ transplants are known to increase cancer risk, according to the National Cancer Institute.

What is KRAS and why do you target it?

For this new study, a team of researchers led by the University of Toronto and Insilico Medicine have created a hybrid quantum classical generation AI algorithm that creates small molecules targeting the KRAS (Kirsten Ras Oncogene Homolog) gene. The KRAS gene belongs to a class of genes called oncogenes. Mutations in the oncogene can cause cancer. In the early 1980s, scientists identified the first human oncogene. In 1982, other scientists in the lab of Harvard Medical School researchers Channing J. Dell and Professor Jeffrey Cooper. Their findings have been published The role of Harvey and Kirsten sarcoma viruses in human bladder and lung cancer.

According to the National Cancer Institute, KRAS gene mutations account for an estimated third of all cancers, and are listed in 90% of pancreatic cancers. Surgical Oncology Clinics in North America Doctors Newhoek, Tsai, and Merrick Bernstam. According to Dr. Shabham Pants, a professor at the University of Texas MD Anderson Cancer Center, KRAS gene mutations are also associated with 40% of colorectal cancer and 32% of lung cancer.

Why quantum computing?

The new realm of quantum computing could be exponentially faster than today’s classical computing (binary computing), which uses binary logic and bits (such as zeros and zeros) to process information .

Zhavoronkov said:

“We presented experiments similar to this quantum work that demonstrated that molecules can be synthesized from the first intertwined conditional autoencoder (ECAAE) and the production AI that is functional experimentally. After that, we presented 2019. In 2019, we published a generative tensor reinforcement learning (GentRL) system, synthesizing and testing the first molecules and tested them in mice. It sets the stages of the entire field, and many others did the same thing. I’ve started. This quantum work is similar to the 2018 ECAAE.”

Quantum computers apply quantum mechanics to solve complex problems and use qubits to store and process data. Not only does Qubits have zeros or one state, but they also have zero-weighted combinations.

Zhavoronkov added:

“From 2026-27, many Hyperscalers, including Microsoft and Amazon, will scale Quantum Machines and open them as a service in Quantum. In China, you can already buy time with actual quantum machines as commercial services. You can also monitor these groups closely to ensure that the software is tailored to Quantum’s drug discovery applications.”

Hybrid Quantum Class AI

Researchers have created a quantum-assisted AI algorithm with a classic computing LSTM algorithm with a quantum-generated AI model. The team generated training datasets of over 1.1 million molecules using 250,000 molecules screened from 100 million molecules using VirtualFlow and 650 experimentally validated KRAS inhibitors. Researchers used hybrid AI models to generate new candidate molecules for targeting KRAS and Insilico Medicine Generation AI Engine Chemistry 42, predicting the top 15 candidates to evaluate in the lab. Researchers discovered that their hybrid quantum classical models are superior to classical models, and that their findings are important proof of concepts.

“Our hybrid quantum model was superior to traditional machine learning models in producing structurally diverse and synthetic drug-like compounds, but 21.5% higher when meeting drug design criteria. There is a success rate, but in this paper we don’t claim to be faster., cheaper than GPU, or better, we just show that it is possible.” says Zhavoronkov.

Copyright©2025 Cami Rosso. Unauthorized reproduction is prohibited.