New Drug Targets in Cancer Treatment Discovered Through Machine Learning

Cancer results from numerous genetic and epigenetic abnormalities. Each patient has tumors with different genetic structures. Precision medicine is developed with the main aim of analyzing these genetic differences to provide the most appropriate treatment. Recently, artificial intelligence and machine learning have increasingly been used in research into new-generation cancer drugs. In April 2018, Cell published a study reporting the discovery of new mutations and drug targets in the Ras protein through machine learning.

In the figure above, you can see definitions and differences among 3 frequently used, closely related computing terms: artificial intelligence, machine learning and deep learning.

Ras Mutations Occur in %30 of All Cancers

Researchers aimed to analyze abnormal protein activities in tumors using machine learning. A branch of artificial intelligence, machine learning used various programs to create prediction models of these protein activities. The model was applied to Ras proteins, which frequently undergo mutations in cancer. Ras proteins are important molecules responsible for cell proliferation and death. In %30 of all cancers, Ras mutations occur. Such mutations cannot be targeted with any drug. The models used data from the Cancer Genome Projectto investigate which drugs the Ras protein would or would not respond to.

According to the research results;

New "hidden" mutations in the Ras protein were identified, along with drugs that might work against these targets.

What Does This Study Mean for Us?

• Cancer has an extremely complex genetic structure. Personalized approaches (precision medicine) are therefore described as the cancer treatments of the future.
• Cancer drugs do not have the same effect in every patient because each has tumors with different genetic structures.
• The Cancer Genome Project aims to map tumors by analyzing genes in actual patient samples. Every gene identified will contribute to treatment planning and help patients receive the most appropriate treatments.
• This study shows that precision medicine and machine learning will revolutionize cancer research. By analyzing thousands of genetic data through machine learning, new-generation cancer biomarkers and drugs can be designed.
• We are living in a golden age of genetics in cancer treatment. CRISPR, the living cancer drug CAR-T and similar new-generation genetic treatments effectively target gene mutations that cause cancer. It must be emphasized that these approaches are still being tested in clinical studies .

Modern medicine continues offering solutions for cancer through contributions from different sciences. There is no single miracle molecule in cancer treatment. Cancer cannot be treated with non-evidence-based herbal treatments (phytotherapy). It can only be treated with modern cancer medicines developed through advanced genetic techniques and proven safe through carefully conducted clinical studies.

RELATED TOPICS

• What Is Machine Learning? Are We Ready for Artificial Intelligence in Healthcare?
• The Limits of the Human Mind, Artificial Intelligence and the Future of Medicine
• Computers in Cancer Diagnosis—Artificial Intelligence Competes with Doctors
• You Cannot Manage What You Do Not Measure—The Surprising History of Health Records!
• How Does Big Data Help Discover New Treatments?