Our work focuses on understanding molecular and clinical aspects of carcinogenesis from a systems biology perspective. Using genomics technologies such as DNA microarrays, my laboratory investigates the transcriptional dynamics and genomic architectures of primary tumors and cell lines at various stages of the oncogenic process and in different clinical contexts. Integrative analysis of gene expression patterns, copy number alterations, and clinicopathologic features allows us to define transcriptional programs of mechanistic and clinical relevance. This strategy has led to the identification and validation of gene expression signatures in liver, breast, ovarian and lung cancers that 1) reflect the activity of specific growth-regulating pathways, 2) define known and novel tumor subtypes, and 3) predict clinical outcomes such as disease recurrence and therapeutic response. Recent examples include prognostic signatures in breast cancer that reflect the operational configuration of the TP53 pathway (Miller et al, PNAS, 2005) and delineate new clinical tumor subtypes based on “genetic grade” (Ivshina et al, Cancer Res, 2006). We have also pioneered novel data mining strategies that integrate several forms of clinico-genomic information (expression, copy number, patient survival) capable of pinpointing known and candidate oncogenes. Using this approach, we have recently identified and validated a novel breast cancer oncogene at chromosome 8p11 that promotes transformation, anchorage-independent growth, invasion through matrigel, and tumor formation in mouse xenograft models. Other candidate genes identified by this method are currently being prioritized for functional validation based on their potential for therapeutic targeting. |