We introduce our incremental coreference resolution system for the BioNLP 2011 Shared Task on Protein/Gene Znteraction. The beneﬁts of an incremental architecture over a mentionpair model are: a reduction of the number of candidate pairs, a means to overcome the problem of underspeciﬁed items in pair-wise classiﬁcation and the natural integration of global constraints such as transitivity. A ﬁltering system takes into account speciﬁc features of different anaphora types. We do not apply Machine Learning, instead the system classiﬁes with an empirically derived salience measure based on the dependency labels of the true mentions. The OntoGene pipeline is used for preprocessing.