Page Loader
IIT Madras's AI tool can predict cancer-causing genes: Here's how
Cancer accounted for nearly 10 million deaths in 2020 (Photo credit: Adobe stock)

IIT Madras's AI tool can predict cancer-causing genes: Here's how

Jul 07, 2022
03:38 pm

What's the story

Researchers at the Indian Institute of Technology Madras (IIT Madras) have found a way to improve the efficiency of cancer treatment. The AI-based mapping tool dubbed 'PIVOT' can predict genes that are responsible for causing cancer. The tool will help in devising personalized strategies for tackling cancer in individuals. According to the World Health Organization, cancer is the leading cause of death worldwide.

Context

Why does this story matter?

The best way to battle cancer is to identify it at an early stage. However, many a time, the disease is detected a tad bit late, which leads to several complications. Even if its detected and treatment begins, current cancer treatments are detrimental to the overall health of the patient. A discovery like PIVOT provides a two-fold solution: much early detection and personalized treatment.

The tool

PIVOT can classify genes based on their cancer-causing ability

PIVOT is designed to predict cancer-causing genes in individuals. It is based on a machine-learning model that can classify genes into oncogenes, tumor suppressing genes, or neutral genes. The tool makes its prediction by utilizing information on mutation, expression of genes, and copy number variation in genes as well as perturbations in the biological network due to an altered gene expression.

Information

The team that made the important discovery

The research was led by Prof. Raghunathan Rengaswamy, Dean (Global Engagement), IIT Madras, Dr. Karthik Raman, core member of Robert Bosch Centre for Data Science and Artificial Intelligence (RBCDSAI) at IIT Madras, and Malvika Sudhakar, a research scholar at IIT Madras.

Importance

IIT Madras study is the first to use supervised learning

There are already tools in existence that can identify personalized cancer genes, but they use unsupervised learning and their predictions are based on the presence/absence of mutations in cancer-related genes. On the other hand, the IIT Madras study uses supervised learning for the first time, and the predictions are made taking into account the functional impact of mutations. This makes predictions more precise.

Information

PIVOT can predict genes related to three kinds of cancer

PIVOT can predict three kinds of cancers at the moment, including Breast Invasive Carcinoma, Colon Adenocarcinoma, and Lung Adenocarcinoma. The researchers plan to add more types of cancers to the list.

Official words

PIVOT will push boundaries of precision medicine

"As cancer treatment increasingly shifts toward personalized medicine, such models that build toward pin-pointing differences between patients can be very useful," said Dr. Raman. "Research in precision medicine is still at a nascent stage. PIVOT helps push these boundaries and presents prospects for experimental research based on the genes identified," said Sudhakar.