Given the amount of sequence data available today, in silico function prediction, which often includes detecting distant evolutionary relationships, requires sophisticated bioinformatic workflows. The algorithms behind these workflows exhibit complex data structures; they need the ability to spawn subtasks and tend to demand large amounts of resources. Performing sequence analytic tasks by manually invoking individual function prediction algorithms having to transform between differing input and output formats has become increasingly obsolete. After a period of linking individual predictors using ad hoc scripts, a number of integrated platforms are finally emerging. We present the ANNOTATOR software environment as an advanced example of such a platform.