Your quest may be very challenging then. Regression analysis may get you in the ballpark but remember that the work in papers mentioned above were based on the rocks they were analyzing. My experience is that every basin and depositional system is different and require basin specific regressions.
Following Diane idea, most of log technics developed require to have at least a well with %TOC analysis from rock samples measured in the Laboratory (the most apropiated samples are core and side wall samples due to its more reliable certainty on real depth). This is necessary for technic calibration.