We are an interdisciplinary group from UC Berkeley and UCSF working in computational pathology. A major focus of the group is on the vast amounts of high-resolution spatial-omics data that are starting to be generated in Pathology laboratories such as digitized H&E stained slides, multiplex immunofluorescence, spatially resolved transcriptomics, and 3D histology data. The scale and complexity of these data (e.g. whole slide images are typically 100,000 x 100,000 pixels and take years of medical training to interpret) mean new statistical/machine learning algorithms need to be developed to fully harness the promise of these emerging tissue measurement technologies.

Linking these different tissue-based data modalities together with other molecular and electronic health record information presents significant opportunities for discovery, but poses major statistical and computational challenges. The computational arm of our work weaves together high-dimensional statistics, network analysis, deep learning, computer vision, and open-source software to solve these challenging biomedical data analysis problems. The scientific arm of our work aims to improve clinical practice through the use of data-driven methods to understand disease development, discover biomarkers, improve diagnosis, and predict response to therapeutics.

Team

Iain Carmichael
Neyman Visiting Assistant Professor, Department of Statistics, UC Berkeley
Adjunct Visiting Assistant Professor, Department of Pathology, UCSF

Huong