In a new article out today in npj Genomic Medicine led by TIME Lab graduate student Maksym Zarodniuk, we use computational analysis, unbiased bioinformatics approaches, and machine learning techniques to stratify brain tumors based on their expression of genes related to extracellular matrix (ECM). Tumors with high ECM expression are found to be enriched with a rare, recently discovered population of perivascular fibroblasts from the brain parenchyma. The presence of this population in brain tumors such as glioblastoma predicts poor response to immunotherapy, as well as upregulation of immune evasive mechanisms such as cancer cell stemness and immunosuppressive myeloid cell activity. This project was performed in collaboration with undergraduate researcher and Grand Challenges Scholar Xander Steele (from the Department of Electrical Engineering), and Professors Xin Lu and Jun Li from the Departments of Biological Sciences, and Applied and Computaitonal Mathematics and Statistics, respectively.