For a lot of separations it would be very helpful to know the value of pK for the analites. Is there any method to calculate these values as long you can not find them in the literature?

I have heard that there are various commercial and proprietary programs for estimating values of pKa from compound structure, but I have also heard that the reliability of these programs can be poor for certain classes of compounds. One such program is pKalc from Compudrug in Budapest, Hungary. We have briefly tested the latter in our lab, and it works more or less well. I presume that you want these values as an aid in method developent, in order to select pH values that will either give you a robust method (no change in separation with pH) or you want to use pH to optimize the separation. In either case, a few measurements with different values of pH can be used to estimate the pKa value for the mobile phase you are using (apparent pKa values vary with %B). Thus, if you plot values of k vs pH, you will obtain a sigmoidal curve similar to a pH titration, and from this you can get pKa = pH at the middle of the curve. Of course, this works best for compounds with pKa values between 2 and 8. Otherwise, you must worry about column stability at extremes of pH.