A new hybrid semantic similarity computation method based on gene ontology

Most existing methods used for computing semantic similarity don´t take full consideration of related factors, therefore they not only fail to handle identical annotations but also show a strong bias toward well-annotated gene or gene products. Concerning these problems, we proposed a new hybrid method based on multiple factors that affect the semantic similarity of Gene Ontology (GO) terms. The new method integrated information content and the structure of GO to compute the semantic similarity of GO terms, which overcomes some serious drawbacks of pure node-based methods and edge-based methods. Experimental results demonstrate that the new method has high accuracy.