Training and testing dataset using Linear Model of machine learning

In the example and the functions provided for the Linear Model in Matlab for example

mdl = LinearModel.fit(X,Y)

mdl contains functions such as MSE which would calculate the error of the "training"set. Is there any similar function for a suitable"testing" set. I am asking this because I am aware of the predict function, and I just would have to write extra amount of code to find the MSE of a testing set. Moreover the residuals of a testing set would not be available in the workspace, when I use predict function. Is there any way which would enable to write less amount of code and find testing error i.e MSE of the Linear Model. Thanks.