Research in my lab involves two interwoven elements: improving the interpretability of network analysis and characterizing transcriptional data in the brain. These topics form a naturally complementary unit because the complexity of the brain as a system means that it is essential that the methods for analyzing it yield clear and precise signals. A dominant interest within computational biology is the analysis of gene networks to provide insight into diverse levels of functional activity, typically starting with regulatory interactions and moving up to more diffuse associations important for understanding systemic dynamics. Gene associations (of various sorts) are believed to encode functional interaction, and this interaction is frequently shown to be able to substantially predict gene function. In this talk, I will demonstrate the power of gene networks determined from shared expression profiles to understand phenotype. I will focus on an a number of unusually well-defined cases, including individual neurons, rare diseases, (outbred) genetic replicates, and largescale populations. In each case, using the observed co-expression of genes to filter complex data for its most important functional properties will yield clear signals that precisely characterize phenotype.