Microscopic image of breast cancer cells resisting treatment. Credit: Callista Images/Image Source/Getty Images
Scientists are increasingly using Artificial Intelligence (AI) and machine learning (ML) algorithms to analyse genomics data and for predictive modelling and precision oncology.
A recent conference on genomic analysis and technology at Delhi’s International Centre for Genetic Engineering and Biotechnology (ICGEB) discussed new advances in this field.
“AI and ML are making it possible to massively improve the accuracy, efficiency and timeliness of cancer detection,” said Dinesh Gupta, group leader of translational bioinformatics at ICGEB. This directly leads to improved patient outcomes and reduced healthcare costs. Additionally, these technologies can take cancer care to underserved populations, he said.
The All India Institute of Medical Sciences (AIIMS) in Delhi recently launched an AI system trained on 500,000 radiological and histopathological images from 1,500 breast and ovarian cancer cases, the most prevalent forms of cancer in India.
Melissa Fullwood, associate professor at Nanyang Technological University in Singapore, shared her team’s use of AI to predict chromatin interactions in cancer. Fullwood’s team developed the Chromatin Interaction Neural Network (ChINN), a convolutional neural network that predicts chromatin interactions using DNA sequences. This AI method aids drug target identification and discovery by analysing genomics and epigenomics data.
New AI-ML tools are now available for cancer pathology and histology analysis, imaging results, and circulating tumour nucleic acids analysis, advancing early cancer detection and precision medicine. Gupta emphasized the role of machine learning in multi-omics, combining data from genomes, transcriptomes, proteomes, epigenomes, and metabolomes for a comprehensive understanding of cancer biology.
Researchers at the Indian Institute of Technology, Dharwad, are using multi-omics-based classification, which proved to be a more accurate prediction model, to identify five novel lung cancer cell clusters with different genetic and clinical features1.
Scientists from the Institute of Bioinformatics and Biotechnology, Manipal, and Kidwai Cancer Institute of Oncology, Bangalore, are also using multi-omics to report different population-specific molecular signatures of ovarian cancer2.
Chad Creighton, professor of medicine at Baylor College of Medicine, noted the importance of proteomics in integrating multi-omics data for a complete molecular picture of cancer. Proteomics can reveal cancer molecular subtypes not identified by transcriptomics, helping identify potential gene targets, he said.
The combination of AI and multi-omics data is advancing precision medicine in cancer. Next-generation sequencing technologies provide specific insights for personalized cancer treatment, enabling oncologists to design targeted therapies based on a comprehensive understanding of genetic mutations and cancer subtypes.
