The analysis of mouse mutants is a powerful tool for studying gene function. Assessment of phenotype is often carried out on mice where littermates of different genotypes are housed together, and this is likely to increase as direct assessment of phenotype in the home cage becomes more common. This raises the important question, particularly for behavioural phenotypes, of the influence of mutant mice on the phenotype of wild-type littermates or vice versa. An interesting study by Kalbassi and colleagues looks at this very phenomenon for mice lacking neuroligin-3, a gene which impacts on social interactions. They show that male neuroligin-3 knockout and wild-type mice when maintained in mixed genotype housing modify each other’s behaviour. They also found that re-expression of neuroligin-3 in mutant mice rescued their social behaviour defects and alleviated the phenotype effects on the wild-type littermates. The genotype interactions are sexually dimorphic, so that the behaviour of female mutants was not affected by their wild-type littermates, but they did modify the behaviour of the co-housed wild-types. There is an increasing realisation that we need to be aware of genotype interactions in co-housed animals, and this important paper exemplifies the care that is needed to separate gene-environment interactions from gene effects in mouse phenotype studies.