Fix for bug #67104 solving the following problems:
- Proper loading of reference tables and solving of a mapping problem (related to bug 66306 where the array now have the size as before but the mapping has changed to start from 0)
- TRD out from combined PID since we have only reference tables for electrons and pions
- Gain normalization factors to match the dEdx in simulation and real data to the gain values in the reference data coming from the testbeam