The process of shear-induced aggregation of fully destabilized colloidal suspensions has been investigated by adopting a mixed deterministicstochastic modelling method. This method is based on a combination of a Monte Carlo algorithm, used to solve in a stochastic way the population balance equation for a purely aggregating suspension, and a Discrete Element Method, employed to simulate aggregation events in a fully predictive manner. The DEM was built in the framework of the well-established Stokesian dynamics technique to get an accurate prediction of the hydrodynamic forces acting on the primary particles. Two different approaches were instead used to describe colloidal interactions: the first assumes primary particles to interact only by means of central forces; a second model assumes also tangential interactions to act on primary particles upon contact. To describe such interactions we adopted a spring-like force model recently proposed by Becker and Briesen. Simulations were performed to ascertain the effect of these two different modelling approaches on the process of aggregation, showing that substantial differences appear in the predicted cluster morphology.