New Delhi, August 30
A team of Indian scientists is making global waves to develop an AI-based system that can select cancer-causing compounds from a chemical crowd, offering huge potential to change the face of the cancer discovery landscape.
The paper, “Artificial Intelligence Reveals Human Cancer-Causing Metabolites,” published in the internationally acclaimed “Nature Chemical Biology,” presents a new AI approach that can identify carcinogens from among a sea of molecules.
The work of four collaborating institutes – Indraprastha Institute of Information Technology Delhi, IIT Ropar, CSIR Institute of Genomics and Integrative Biology and Rajiv Gandhi Cancer Institute, has already been hailed as a game-changer in the field of oncology. Percent of cancers are hereditary and the rest are caused by exposure to carcinogens.
“Since more than 90 percent of cancers are caused by exposure to compounds, the ability of our AI method to accurately identify a carcinogen can make a huge difference in response to cancer. Over the past few decades, more than 700 drugs approved by the Food and Drug Administration because they were later found to be carcinogenic even though they passed the trial stages. It is a huge burden for drugs to determine the potential for carcinogenicity of a compound. Gaurav Ahuja, senior author of the study, told the Tribune on Tuesday that the AI system Our Metabokiller is a reliable method for accurate detection of carcinogenic compounds.
Indian researchers have demonstrated the value of their AI model by identifying two compounds (none of the cancer-detection models currently in use indicate they cause cancer) as causing cancer and demonstrating this in experiments.
Ahuja and collaborator Dr. Debarka Sengupta say current predictive models for predicting cancer primarily examine compounds to see if they cause DNA damage and predict cancer based on this knowledge.
“But a potential carcinogenic compound may not primarily damage DNA alone. Carcinogens can have many other biochemical properties that our AI model evaluates for their detection. It greatly expands the ability to detect carcinogens. It can be used by the cosmetic, food, beverage and pharmaceutical industries alike,” says Ahuja.
Metabokiller follows biochemical properties known to be associated with carcinogens to detect cancer compounds – a potential carcinogen may lead to cell proliferation (cell division), genetic instability, anti-apoptotic response (the ability of the cell not to die), and epigenetic modifications (changes to DNA) ) and may have an electrophilic character (it induces a positive charge to affect negatively charged DNA). The AI program has already outperformed all current carcinogenicity detection models.
When asked how this would help, Ahuja explains, “Pharma companies spend billions to check whether their new drugs can cause cancer. The process takes at least six years and includes animal studies. It will enable carcinogens to be detected on a laptop in two seconds. Compounds tagged through our program can then enter priority research by pharmaceutical companies which can save money, time and animal lives.”
IIITD researchers are now working on a predictive model to detect any carcinogenic attacks on the gene present in the human body. “Gall bladder cancer is common among people who live in the Ganges belt. Cancers are often associated with the region people live in. We are now working to establish a connection between the carcinogen and the gene it targets and turns on,” says Ahuja.