Our team is focused on using integrative approaches to genomic data to identify genetic biomarkers for tumor progression and response to therapy. While many genomic analyses in cancer have focused on a single type of genomic data, we are interested in integrating somatic mutations and gene expression profiles with germline genetic information to build rigorous predictive models. For example, variants (germline or somatic) associated with the immune state of the tumor microenvironment may inform our understanding of response to immune-oncology therapies. We work in a highly multidisciplinary environment and blend rigorous statistical approaches and bioinformatics with detailed biological understanding.
Our studies are powered by the wealth of genomic and phenotypic data available in public initiatives such as The Cancer Genome Atlas (TCGA) as well as data from internal cancer cell line screens. Given that these datasets have millions of variants and tens of thousands of genes assessed in much smaller numbers of clinical samples, we are interested in developing and applying statistical approaches to high-dimensional data, including methods for feature space reduction (e.g., focusing on frequently mutated genes, immune-relevant pathways, or genes with validated phenotypic associations) to identify the most impactful variants and genes.