© 2015 American Chemical Society. In nanopore sensing experiments, the properties of molecules are probed by the variation of ionic currents flowing through the nanopore. In this context, the electronic properties and the single-layer thickness of graphene constitute a major advantage for molecule characterization. Here we analyze the translocation pathway of the thioredoxin protein across a graphene nanopore, and the related ionic currents, by integrating two nonequilibrium molecular dynamics methods with a bioinformatic structural analysis. To obtain a qualitative picture of the translocation process and to identify salient features we performed unsupervised structural clustering on translocation conformations. This allowed us to identify some specific and robust translocation intermediates, characterized by significantly different ionic current flows. We found that the ion current strictly anticorrelates with the amount of pore occupancy by thioredoxin residues, providing a putative explanation of the multilevel current scenario observed in recently published translocation experiments.