Coreference resolution is the task of identifying the sets of mentions referring to the same entity. Although modern machine learning approaches to coreference resolution exploit a variety of semantic information, the literature on the effect of relational information on coreference is still very limited. In this paper, we discuss and compare two methods for incorporating relational information into a coreference resolver. One approach is to use a filtering algorithm to rerank the output of coreference hypotheses. The filter is based on the relational structures between mentions and their corresponding relationships. The second approach is to use a joint model enriched with a set of relational features derived from semantic relations of each mention. Both methods have shown to improve the performance of a learning-based state-of-the-art coreference resolver.