plsregress for new data set and for different PC

Hi,

I would like to know how to predict Y for different X and different principal components without running always plsregress. I have tried the following:

[XL,YL,XS,YS,BETA,PCTVAR,MSE,stats]=plsregress(X1,Y1,PCmax);

Y1=[ones(size(X1,1),1) X1]*BETA;

XS2=X2*stats.W(:,PCi);

Y2=XS2*YL(:,PCi)';

And for X1=X2 I obtained different Y1 and Y2.

Any thoughts to sort out this issue?

Thank you.

Cheers,

Joao