In the last two decades, the scope and scale of biomedical data have expanded to capture the holistic definition of health – “a state of complete physical, mental and social well-being and not merely the absence of disease or infirmity (WHO)”. Deciphering the complex patterns embedded in this high-dimensional data holds great potential to accelerate the integration of personalized medicine into clinical workflows. With its ability to learn representations directly from the data, Machine Learning (ML) has become an integral part of analytical frame works across diverse domains of health research, including disease-risk stratification, early diagnosis, prognosis prediction, evidence-based clinical decision-making, and the development of personalized therapeutic strategies. Though the accumulating research continues to highlight the transformative potential of ML in healthcare, majority of these ML discoveries fail to generalize in real-world settings. The limited success of ML models in real-world settings could be majorly attributed to one or more of the following issues: non-uniform data representation (heterogeneity), sampling bias and overfitting (limited generalizability).  

In this talk, Dr Bhatraju will share insights from his experience applying ML techniques to breathomics data—complex, high-dimensional data capturing volatile organic compounds from exhaled breath—for the classification of chronic lung diseases such as asthma, chronic obstructive pulmonary disease (COPD), and lung cancer. He will discuss the common data challenges that limit the clinical usefulness of ML models used to decode the breathomics data. The goal is to underscore the importance of domain knowledge and contextual understanding of data in guiding thoughtful ML design and validation to improve the generalizability of the developed models and thus enhance the real-world healthcare impact.

Dr Naveen Kumar Bhatraju is a Research Faculty Fellow in the Centre for Health Analytics Research and Trends (CHART), and the Koita Centre for Digital Health (KCDH-A) at Trivedi School of Biosciences (TSB), Ashoka University. He earned his PhD in Biophysics from the University of Hyderabad and pursued postdoctoral research at the CSIR–Institute of Genomics and Integrative Biology, New Delhi. Dr Bhatraju’s research lies at the intersection of computational biology and digital health. His work focuses on uncovering how genetic, lifestyle, and environmental factors interact to shape the individual health trajectories and influence the risk of age-related diseases. By applying data-driven and machine-learning approaches, he aims to advance the understanding of complex health dynamics and translate these findings into personalized healthcare strategies.

Mandatory for students pursuing DATA204.