Revealing the structure and the intrinsic dynamics of a protein is paramount to understand its function and other properties of evolutionary importance. Several schemes to obtain such insight in silico were developed over the decades. A computationally efficient protocol approximates the molecular dynamics around its native state by a harmonic potential. In this paper, we introduce a new methodology to combine the various harmonic approaches to understand the folding/unfolding dynamics and the dynamics around the native structure of the protein in a temperature dependent way. We apply this new protocol to the HIV1-protease and discuss the results in the light of events in the adaptive evolution towards drug resistance, which is a major problem in HIV infection.