A new method was developed to extract relationships between biological entities using a combination of different kernel learning techniques. By combining two separate methods into one, the performance of the new method was found to be better than using each method individually. This approach was tested on five different evaluation datasets, showing promising results in extracting biological entity relationships.