We give an overview of one of the three downstream systems in the Extrinsic Parser Evaluation shared task of 2017: the Trento--Gothenburg system for opinion extraction. We describe the modifications required to make the system agnostic to its input dependency representation, and discuss how the input affects the various submodules of the system. The results of the EPE shared task are presented and discussed, and to get a more detailed understanding of the effects of the dependencies we run two of the submodules separately. The results suggest that the module where the effects are strongest is the opinion holder extraction module, which can be explained by the fact that this module uses several dependency-based features. For the other modules, the effects are hard to measure.