The number of potential phenotypes derived for human neurogenetics experimentation is almost infinite, including measurements within single neurons (and/or glia), between local populations of neurons, between systems of spatially discreet networks, or from observable behavior. Indeed, even experiments focused exclusively on functional or structural neuroimaging methods, which are typically restricted to assessing system level measures, can produces hundreds of thousands to millions of phenotypes. Thus, one significant quandary for researches using neuroscientific methods to discover gene involved in brain-related illnesses like Alzheimer’s disease, autism, bipolar disorder, depression, multiple sclerosis, or schizophrenia is how to choose between these potential phenotypes. This issue is compounded when using very large genome arrays or whole genome sequence data. A major focus of my laboratory is the development of empirical methods for reducing the phenotype space prior to genetic inquiry. Recently, we developed a method for ranking potential allied phenotypes for a particular illness on their common genetic co-variation. Today, I will discuss our work with applying neurocognitive, neuroimage and transcriptional phenotypes in search of risk genes for psychotic and affective illnesses utilizing this method. Specifically, I will present work from the “Genetics of Brain Structure and Function” study, which involves acquisition of behavioral, neurocognitive and neuroimaging phenotypes in 1500 Mexican Americans from randomly-selected extended pedigrees. All participants have high-density SNP arrays, transcriptional data from two time points and 502 individuals have 60-fold genome-wide sequence data. Our approach involves localizing loci for a phenotype via genome-wide linkage or association, identifying the non-synonymous or regulatory variants diving that effect with sequence data, and then demonstrating pleiotropy with published association studies. Our results provide clear examples of how to use empirical guidelines to choose phenotypes that can provide novel genetic insights for brain-related illness.