The Genobox technology is build on the principle that the phenotypic properties of a micro-organism can to a large extent be predicted on basis of its genome sequence.

To do these predictions, the Genobox server contains a large number of predictive models. These predictive models are based on known genes with a known function and a known relation to phenotypic properties. Matching the predictive models to the genome sequence for individual organisms yields the phenotypic predictions. This process is schematically depicted below for a simple example consisting of only 2 genes, in this case for the production of cysteine.