Figure 1. Illustration of event extraction and gene normalization.

The gene mentions recognised in text are in red and the extracted event structures in blue. The normalization algorithm further maps the ambiguous gene mentions to unique database identifiers (in green).