University of Virginia researchers are applying machine learning and functional genomics to investigate a largely unexplored region of the genome — transcribed ultra conserved regions (TUCRs) — and their potential role in glioblastoma, one of the most aggressive and deadly forms of brain cancer.

Stephen Turner (School of Data Science) and Roger Abounder (microbiology, immunology, and cancer biology) are leading the collaboration, supported by a pilot grant connecting UVA's Cancer Center and data science researchers. Despite decades of genomic research, the protein-coding portion of the genome — only a small fraction of the total — has received the lion's share of scientific attention. TUCRs, by contrast, have fewer than 70 publications to their name, with none previously examining their role in glioblastoma.

Glioblastoma carries a median survival of just over one year even with surgery, chemotherapy, and radiation combined, underscoring the urgency of new research avenues. Turner and Abounder hope that shining a light on this uncharted genomic territory will yield insights that could ultimately improve outcomes for patients.

The project was presented at a Cancer-Data Science Research Symposium at UVA on February 19, an event designed to connect cancer researchers with data scientists and foster exactly this kind of cross-disciplinary collaboration.

Artificial intelligence (AI) is emerging as a powerful force in breast cancer research, but the future of personalized care will not be written by algorithms alone. 

Rockefeller researchers previously discovered several new important tasks that the antioxidant glutathione manages throughout the body. Now they’ve found that glutathione maintains the smooth operations of a protein-producing hub in the cell called the endoplasmic reticulum by acting as a protein “proofreader.” When this process goes awry, it can lead to neurodegeneration and cancer.