Data Fusion Strategies for Enhancing Classification Performance: a Case Study on Cytochrome P450