The perturbation of a protein conformation by a physiological fluid flow is crucial in various biological processes including blood clotting and bacterial adhesion to human tissues. Investigating such mechanisms by computer simulations is thus of great interest, but it requires development of ad hoc strategies to mimic the complex hydrodynamic interactions acting on the protein from the surrounding flow. In this study, we apply the Lattice Boltzmann Molecular Dynamics (LBMD) technique built on the implicit solvent coarse-grained model for protein Optimized Potential for Efficient peptide structure Prediction (OPEP) and a mesoscopic representation of the fluid solvent, to simulate the unfolding of a small globular cold-shock protein in shear flow and to compare it to the unfolding mechanisms caused either by mechanical or thermal perturbations. We show that each perturbation probes a specific weakness of the protein and causes the disruption of the native fold along different unfolding pathways. Notably, the shear flow and the thermal unfolding exhibit very similar pathways, while because of the directionality of the perturbation, the unfolding under force is quite different. For force and thermal disruption of the native state, the coarse-grained simulations are compared to all-atom simulations in explicit solvent, showing an excellent agreement in the explored unfolding mechanisms. These findings encourage the use of LBMD based on the OPEP model to investigate how a flow can affect the function of larger proteins, for example, in catch-bond systems.