IIITH uses genetics, AI to detect cancer early

Hyderabad: At a time when cancer care is shifting from one-size-fits-all treatment to precision medicine, researchers at the International Institute of Information Technology, Hyderabad, are weaving together genetics, epigenetics and AI to push the frontiers of early detection and personalised therapy. At the Centre for Computational Natural Sciences and Bioinformatics (CCNSB) in IIIT, professor Nita Parekh and her team are integrating genetic mutations, epigenetic changes and patterns in medical images to better understand how cancers arise and behave. "People talked about the origin of tumours since the early 1900s, but over time we realised that cancer cannot be explained by mutations alone. Today, cancer is understood as a multifactorial disease, shaped by genetics, gene regulation, environment, and time," observes professor Parekh.Cancer genomesHer group is analysing genetic variations to map differences across cancers and their subtypes. "One part of my work looks at the role of genetic variations in tumorigenesis," says professor Parekh, referring to both small single letter DNA changes and larger alterations such as deletions, duplications, inversions and gene fusions. "By analysing cancer genomes in detail, we can identify which mutations matter, which pathways they disrupt, and how different cancers—or even subtypes of the same cancer—behave very differently," states professor Parekh. In diffuse large B-cell lymphoma, an aggressive blood cancer with two main subtypes, the team identified subtype-specific mutations, disrupted pathways and biomarkers that explain why some patients respond well to treatment while others do not.Epigenetic regulation"Genes alone do not tell the full story," she says, highlighting ongoing work on DNA methylation. "Epigenetic regulation is the ongoing part of my work." The team is studying regulatory shifts across promoters, enhancers and non-coding RNAs. "These RNAs do not code for proteins, but they play a major role in regulating gene expression," she says.In breast cancer, researchers are combining methylation data, RNA profiles and machine learning to identify early markers and subtype-linked signatures. "Breast cancer is not one disease; it has several subtypes, and identifying them early makes a critical difference," says Prof Parekh.The team is also building AI-driven mammography analysis. "We are also working on mammography data analysis, which we are developing in parallel," notes Prof Parekh, describing curated datasets used to train models for early detection, segmentation, and classification.