Biostatistics Seminar Series - Matthew McCall, PhD

Title: Quantifying Uncertainty in the Estimation of Gene Regulatory Networks Abstract: Given the numerous sources of variability and technical biases inherent in genomic technologies, network estimation algorithms are often unable to accurately reconstruct gene networks. To address this challenge, we propose an approach to network estimation that explicitly models and incorporates uncertainty in each step of the analysis. Instead of attempting to infer a single "best" network, we report a posterior density on the network space that directly conveys the uncertainty in the inferred network structure. Quantifying the uncertainty in specific network features allows researchers to determine highly-probable features and areas in which additional information is needed, thereby guiding future experimentation. We have applied this approach to a network of cooperation response genes (CRGs), which respond synergistically to loss-of-function p53 and Ras activation. CRGs have been shown to play a crucial role in tumor formation independent of the initiating mutations, and many CRGs are essential components of the cellular machinery involved in maintaining malignancy. Ongoing examination of the CRG network architecture has the potential to uncover specific vulnerabilities of the cancer cell and, ultimately, to guide multi-target interventions.