Effectively integrating information content and structural relationship to improve the GO-based similarity measure between proteins