An animal's behavioral repertoire is deeply rooted in its genome. Mutations of behaviorally important genes may alter or disrupt either the physiology of neuronal circuits or their development. The first task of a research program aimed at identifying the genetic underpinnings of perception and behavior is to build a comprehensive catalog of genes with specific, non-lethal phenotypes, initially with no regard of when and where in the organism they are acting. Forward genetic screens are the method of choice to identify those genes in an unbiased fashion. This approach was pioneered over 30 years ago by Benzer in Drosophila melanogaster  and was quickly extended to Caenorhabditis elegans . In these invertebrate species, the forward genetic strategy was particularly productive for the analysis of sensory systems, such as vision, mechanosensation, and olfaction, where these screens helped to discover many genes important for the patterning of sensory epithelia and for sensory transduction [3–7].
Very few behavioral screens have been attempted in vertebrates to date. In mice, Takahashi and colleagues carried out a screen for dominant mutations disrupting circadian behavior . Other groups have carried out behavioral “shelf screens” of previously discovered mutants in both zebrafish and mice [9–11] or collected mutants in motility and locomotor coordination [12,13]. Here we report on the results of the first large-scale behavioral screen focused on a vertebrate sensory system. Following chemical mutagenesis, we searched for recessive mutations that disrupt visually evoked behaviors in zebrafish. Brockerhoff et al. first showed the utility of optokinetic behavior as a powerful screening tool to find visual mutants . Here we used both the optokinetic response (OKR) and the optomotor response (OMR) as screening assays [9,14–16]. These two behaviors employ different motor outputs (swimming and eye movements, respectively), but they are both elicited by large-field motion and are dependent on the retina as the light-sensing organ [15,17]. In a high-throughput screen of almost 2,000 mutagenized genomes, we discovered 41 loci whose mutations lead to a broad spectrum of specific visual (or visuomotor) impairments. Some of the more striking phenotypes include new mutants in retinal axon targeting and in the adaptive dynamics of light responses. This first survey reveals the extent to which single-gene mutations can perturb visual behavior without affecting gross development or vital organ functions. The identities of the corresponding genes are beginning to provide novel insights into how the visual system is assembled and how cellular and molecular interactions shape sensory processing in the vertebrate brain.