Collecting, storing, and mining research data in a brain bank

The Stanley Medical Research Institute Brain Collection distributes samples from specified cohorts that contain demographically matched groups of subjects with mental illnesses such as schizophrenia, bipolar disorder, and major depression, as well as unaffected controls. The groups are matched by age, sex, race, postmortem interval, pH, side of brain, and mRNA quality. The samples are distributed coded so that all data must be returned in order to obtain the demographic information. The database contains more than 5000 individual data sets, as well as data from high-throughput microarray, sequencing, and proteomic studies. While most data were generated from the frontal cortex and hippocampus, the cerebellum has the most data sets that differ significantly between diagnostic groups and controls. The database contains interactive features and statistical tools that enable online data mining and real-time data analysis. The decrease in density of parvalbumin-positive neurons in schizophrenia, one of the most replicated findings in the field, is used to illustrate features of the brain bank. We describe how this finding can be replicated and validated in this database. We also show how the density of parvalbumin-positive neurons is correlated with markers of immune activation in the neuropathology data sets, how it is correlated with immune-related genes in a microarray data set, and how it is associated with a single-nucleotide polymorphism in the immune complement system.