Toward Predictive Models of Cancer: Early Results and Computational Challenges

Abstract: 
Recent advances in high-throughput sequencing have generated unprecedented volumes of transcriptomic data, offering new opportunities to study cancer as a complex, data-driven system. However, translating these data into predictive and interpretable models remains a major computational challenge. In this talk, I will present our ongoing efforts to develop computational frameworks for the analysis of bulk and single-cell RNA sequencing data, with a focus on B-cell malignancies. I will discuss key challenges, including high dimensionality, data heterogeneity, and integration across datasets, and present preliminary results that highlight emerging gene programs and potential clinical associations. Finally, I will outline future directions toward building predictive models of tumor behavior and enabling more precise, data-driven approaches to cancer research.

Short CV:
Luciano Cascione is a bioinformatician, head of the Bioinformatics Core Unit at the Institute of Oncology Research (IOR) in Bellinzona (Switzerland), and Group Leader at the Swiss Institute of Bioinformatics (SIB). His research focuses on computational analysis of transcriptomic data to investigate cancer biology, with a particular interest in B-cell malignancies. He works on integrating bulk and single-cell RNA sequencing data to identify gene regulatory programs and clinically relevant biomarkers. He has developed computational tools for RNA analysis, including methods for circular RNA characterization and transcriptomic profiling. His current work aims at building predictive and interpretable models to better understand tumor heterogeneity and support precision medicine.