Multivariate analysis of variegated expression in Neurons: A strategy for unbiased localization of gene function to candidate brain regions in larval zebrafish.

Behavioral screens in model organisms have greatly facilitated the identification of genes and genetic pathways that regulate defined behaviors. Identifying the neural circuitry via which specific genes function to modify behavior remains a significant challenge in the field. Tissue- and cell type-specific knockout, knockdown, and rescue experiments serve this purpose, yet in zebrafish screening through dozens of candidate cell-type-specific and brain-region specific driver lines for their ability to rescue a mutant phenotype remains a bottleneck. Here we report on an alternative strategy that takes advantage of the variegation often present in Gal4-driven UAS lines to express a rescue construct in a neuronal tissue-specific and variegated manner. We developed and validated a computational pipeline that identifies specific brain regions where expression levels of the variegated rescue construct correlate with rescue of a mutant phenotype, indicating that gene expression levels in these regions may causally influence behavior. We termed this unbiased correlative approach Multivariate Analysis of Variegated Expression in Neurons (MAVEN). The MAVEN strategy advances the user's capacity to quickly identify candidate brain regions where gene function may be relevant to a behavioral phenotype. This allows the user to skip or greatly reduce screening for rescue and proceed to experimental validation of candidate brain regions via genetically targeted approaches. MAVEN thus facilitates identification of brain regions in which specific genes function to regulate larval zebrafish behavior.