The use of antisense morpholinos to perform targeted knockdowns in zebrafish embryos can be a powerful method to assess the developmental functions of specific genes of interest, but the usefulness of this approach is hampered by the potential for off-target effects. To explore this issue further, Nathan Lawson and colleagues (Dev. Cell doi:10.1016/j.devcel.2014.11.018; 18 December 2014) used genome editing techniques to introduce mutations into 24 zebrafish genes suspected to be involved in early vascular development. They found that mutations in only 3 of the 24 genes caused vascular phenotypes, whereas the other mutations produced no overt defects in vascular or overall morphology. Notably, previous morpholino-based studies had reported developmental phenotypes for more than half of these genes. The low concordance rate between morphant and mutant phenotypes observed in this study indicates that morpholinos may potentially produce very high rates of false positive phenotypes. Given the recent development of efficient methods for generating targeted mutations using site-specific endonucleases, the authors recommend that the generation and characterization of mutants be considered the standard approach for defining gene function in zebrafish.