Blab Tech
IIIT Delhis MutAIverse AI Platform Decodes DNA Damage to Trace Smokeless Tobacco-Related Head and Neck Cancer
A team at Indraprastha Institute of Information Technology (IIIT) Delhi has unveiled a generative artificial‑intelligence platform that translates the chemical scars left on DNA into a forensic trail of the carcinogen responsible for head and neck cancer.
The system, dubbed MutAIverse, was tested on 27 patients with head and neck cancer in Guwahati, Assam, and linked the disease to exposure to smokeless tobacco. According to a study published in the Journal of Cheminformatics, MutAIverse is the world’s first generative AI platform that decodes DNA adducts—chemical modifications of DNA that can lead to cancer.
The platform expands the reference library of known DNA adducts from fewer than 400 entries to more than 300,000. By training on this enlarged database, the AI can generate hypotheses about which genotoxic chemicals produced the observed DNA damage.
At the heart of the system is a machine‑learning tool called AdductLinker. When a tumour biopsy is analysed by mass spectrometry, AdductLinker works backward from the detected DNA adducts to the specific environmental chemical or toxin that created them. In the Guwahati study, the algorithm identified chemical signatures that matched those produced by smokeless tobacco products. The output includes a confidence score, a structural representation of the damaged DNA, and a list of intermediate metabolites that may form inside the body.
Researchers from IIIT Delhi collaborated with the National Institute of Pharmaceutical Education and Research (NIPER) and the Council of Scientific and Industrial Research (CSIR) – Institute of Genomics and Integrative Biology (IGIB). DNA adductomics was performed on tumour biopsies collected by the Dr. Bhubaneswar Borooah Cancer Institute. The study found selective enrichment of both known and novel DNA adducts in the samples, supporting the idea that the platform can detect previously uncharacterised chemical‑DNA interactions.
While the results are promising, experts caution that the platform is a research tool and does not yet prove that a particular DNA adduct caused the cancer. Associate professor Saravanan Matheshwaran of IIT Kanpur noted that detecting an adduct only shows that a chemical has interacted with DNA; it does not establish a causal link to mutation or tumour development. He added that additional experimental, clinical and epidemiological studies are needed to confirm causation.
Similarly, computational biologist Kedar Natarajan of Ashoka University said that the platform could be useful for basic research and for building a more complete picture of cancer development when combined with genome sequencing and other approaches. However, he emphasized that it is not a clinical diagnostic tool and cannot be used for hospital or community screening at this time.
The study highlights the potential of AI to add a new layer of information to cancer research by pointing to specific environmental exposures. If validated in larger cohorts, MutAIverse could help identify high‑risk populations and inform public‑health interventions, especially in regions where smokeless tobacco use is common.
At present, the platform remains in the validation phase. The research team plans to test MutAIverse on additional tumour types and larger patient groups. No regulatory approvals or clinical trials have been announced. The next steps will involve further benchmarking against established biomarkers and exploring integration with existing diagnostic workflows.
In summary, IIIT Delhi’s MutAIverse represents a novel application of generative AI to DNA damage analysis, successfully linking head and neck cancer in Guwahati patients to smokeless tobacco exposure. The platform’s expanded DNA adduct library and AdductLinker algorithm offer a promising research tool, but further validation is required before it can be considered for clinical use.
The system, dubbed MutAIverse, was tested on 27 patients with head and neck cancer in Guwahati, Assam, and linked the disease to exposure to smokeless tobacco. According to a study published in the Journal of Cheminformatics, MutAIverse is the world’s first generative AI platform that decodes DNA adducts—chemical modifications of DNA that can lead to cancer.
The platform expands the reference library of known DNA adducts from fewer than 400 entries to more than 300,000. By training on this enlarged database, the AI can generate hypotheses about which genotoxic chemicals produced the observed DNA damage.
At the heart of the system is a machine‑learning tool called AdductLinker. When a tumour biopsy is analysed by mass spectrometry, AdductLinker works backward from the detected DNA adducts to the specific environmental chemical or toxin that created them. In the Guwahati study, the algorithm identified chemical signatures that matched those produced by smokeless tobacco products. The output includes a confidence score, a structural representation of the damaged DNA, and a list of intermediate metabolites that may form inside the body.
Researchers from IIIT Delhi collaborated with the National Institute of Pharmaceutical Education and Research (NIPER) and the Council of Scientific and Industrial Research (CSIR) – Institute of Genomics and Integrative Biology (IGIB). DNA adductomics was performed on tumour biopsies collected by the Dr. Bhubaneswar Borooah Cancer Institute. The study found selective enrichment of both known and novel DNA adducts in the samples, supporting the idea that the platform can detect previously uncharacterised chemical‑DNA interactions.
While the results are promising, experts caution that the platform is a research tool and does not yet prove that a particular DNA adduct caused the cancer. Associate professor Saravanan Matheshwaran of IIT Kanpur noted that detecting an adduct only shows that a chemical has interacted with DNA; it does not establish a causal link to mutation or tumour development. He added that additional experimental, clinical and epidemiological studies are needed to confirm causation.
Similarly, computational biologist Kedar Natarajan of Ashoka University said that the platform could be useful for basic research and for building a more complete picture of cancer development when combined with genome sequencing and other approaches. However, he emphasized that it is not a clinical diagnostic tool and cannot be used for hospital or community screening at this time.
The study highlights the potential of AI to add a new layer of information to cancer research by pointing to specific environmental exposures. If validated in larger cohorts, MutAIverse could help identify high‑risk populations and inform public‑health interventions, especially in regions where smokeless tobacco use is common.
At present, the platform remains in the validation phase. The research team plans to test MutAIverse on additional tumour types and larger patient groups. No regulatory approvals or clinical trials have been announced. The next steps will involve further benchmarking against established biomarkers and exploring integration with existing diagnostic workflows.
In summary, IIIT Delhi’s MutAIverse represents a novel application of generative AI to DNA damage analysis, successfully linking head and neck cancer in Guwahati patients to smokeless tobacco exposure. The platform’s expanded DNA adduct library and AdductLinker algorithm offer a promising research tool, but further validation is required before it can be considered for clinical use.
