Breast cancer remains the most commonly diagnosed cancer among women worldwide. According to the World Health Organization (WHO), approximately 2.3 million women were diagnosed with breast cancer globally in 2022, with about 670,000 deaths.
Bangladeshi researcher Umme Habiba, a PhD candidate in Mathematical and Statistical Sciences at the University of Texas Rio Grande Valley (UTRGV), USA, is developing an artificial intelligence model aimed at improving drug response prediction in breast cancer treatment.
Her research integrates advanced mathematics, multi-omics biological data, and transformer-based deep learning to analyze complex genetic information related to cancer.
Habiba’s earlier research focused on nonlinear fractional partial differential equations (PDEs), where fractional operators were used to model systems with nonlocal and memory-dependent dynamics. This mathematical background naturally supports her current work in machine learning for biomedical systems, where large-scale biological data also involve complex nonlinear interactions. In both areas, her work focuses on designing structured mathematical representations that simplify complex systems.
Building on this foundation, her current research maps gene-level biological information to known biological pathways, and then groups related pathways into higher-level structures called super-pathways, representing coordinated cellular processes. These super-pathways serve as structured tokens in a transformer-based artificial intelligence model used to predict cancer subtypes and drug response patterns. The framework can also incorporate PDE-inspired constraints as part of the training loss function, integrating mathematical modeling with modern deep learning approaches.
Modern cancer research increasingly relies on large-scale genomic and molecular data to understand why patients respond differently to the same treatment. However, interpreting these complex datasets remains a major scientific challenge. Advanced computational models that can analyze multi-omics data and uncover meaningful biological patterns are becoming essential tools for improving cancer treatment strategies.
By combining mathematical modeling with artificial intelligence, Habiba’s research aims to develop more interpretable and reliable predictive systems for biomedical applications. Such approaches could help researchers better understand the biological mechanisms behind drug response and support the development of more personalized cancer therapies in the future.
Researchers believe that data-driven artificial intelligence systems, such as the one being developed by Habiba, could contribute to these global efforts by helping physicians better understand treatment response and move toward more effective precision oncology strategies.