I am trying to compute scores - I want to find the principal componet in a set of spectra. So for each observation I may have 700 absorbancies and several hundred observations. So I want to find the optimal number of componets to explain my variance. Later on I will build a prediction model using MLR or PLS. Typical Y values in my matrix when I build the model will be constituents like Moisure, Fat and so on. But initially I just want to focus on the X variables - the 700 absorbancies per observation or sample. I hope this is the type info you were looking for.