The task of event extraction in the biomedical domain corresponds to the systematic identification of interactions between different biomolecule entities in text. The biomedical community has been working towards the goal of creating a curated knowledge base of biomolecule entity interactions. Scientific literature in biomedical domain containing millions of articles is an excellent source of gathering such information. However automatically extracting information from text is a challenging task since natural language allows us to express interactions between entities in various different ways. Current approaches to this problem span from methods that use manually crafted rules to more machine learning methods that use syntactic parses as features. However a more semantic analysis of text that tries to normalize such variations in text could be more useful. In this talk I will describe our current ongoing work where we make use of one such semantic representation called the Abstract Meaning Representation (AMR) (Banarescu et al., 2013). AMR is a rooted, directed acyclic graph (DAG) that captures the notion of "who did what to whom" in text in a way that sentences that have the same basic meaning would have the same AMR. We cast the event extraction problem as the problem of identifying a subgraph from the AMR graph. We train a neural network based model that encodes the path information in the AMR to identify an event subgraph. Finally, we test our hypothesis on the BioNLP Shared Task dataset.