Changes in the relative biological effectiveness (RBE) of the radiation-induced cell killing of human salivary glands (HSG) were assessed along the Bragg peak of a 60 MeV clinical proton beam by means of coupling biophysical models with the results of Monte Carlo radiation transport simulations and experimental measurements with luminescent detectors. The fluence- and dose-mean unrestricted proton LET were determined along the Bragg peak using a recently developed methodology based on the combination of the response of 7LiF:Mg,Ti (MTS-7) and 7LiF:Mg,Cu,P (MCP-7) thermoluminescent detectors. The experimentally assessed LET values were compared with the results of radiation transport simulations using the Monte Carlo code PHITS, showing a good agreement. The cell survival probabilities and RBE were then calculated using the linear-quadratic model with the linear term derived using a phenomenological LET-based model (Carabe A et al 2012 Phys. Med. Biol. 57 1159) in combination with the experimentally-assessed or PHITS-simulated dose mean proton LET values. To the same aim, PHITS simulated microdosimetric spectra were used as input to the modified microdosimetric kinetic model (modified MKM, (Kase et al 2006 Radiat. Res. 166 629-38)). The RBE values calculated with the three aforementioned approaches were compared and found to be in very good agreement between each other, proving that by using dedicated pairs of thermoluminescent detectors it is possible to determine ionization density quantities of therapeutic proton beams which can be applied to predict the local value of the RBE.