how to combine the different features vectors to make single feature vector?

Hi everybody, Can anyone tell me how to combine the different features vectors to make single feature vector? I find that integrating features from different sources using a Multiple-Kernel-Learning (MKL) approach associated with a kernel-based classifier such as SVM has been reported to be effective. How can I combine them by MKL in matlab? Is there built-in function? Is there any suggestion to combine multiple different size feature vectors?