The modelling of the behaviour of crude oils is particularly challenging due to the diverse length and time scales involved, and the inherent uncertainty concerning the chemical characterization of the system. Here we describe a framework that spans multiple time, size and complexity scales to provide a richer description of the asphaltene aggregation process. At the bottom of the scale, the Quantitative Molecular Representation (QMR) approach was used to generate a description of plausible molecular structures of asphaltenes and resins based on experimental data from crude oils, consisting of elemental analysis and both 1H and 13C NMR spectroscopy parameters. The behaviour of the QMR-generated model asphaltene structures in heptane and toluene solvents was then determined using a fully atomistic description in classical molecular dynamics (MD) simulations, shedding light upon the incipient clustering mechanisms and the size and distribution of asphaltenes in simple solvents. While simulations at this level of detail are enlightening, current computational resources are insufficient to cover the times required to represent the clustering process. Appropriately, at the next level of scale, coarse-grained (CG) models were built using segments representing multiple atoms, significantly reducing the number of particles and calculations required in the simulation. We used here the Statistical Associating Fluid Theory force field, where the parameters describing the intermolecular potential are obtained from macroscopic thermophysical properties. From this we have produced CG models of asphaltenes, resins and other saturated and aromatic components typical of a live oil. Complex systems composing of realistic mixtures of solvents, resins and multiple asphaltenes in wide ranges of concentrations, pressures and temperatures were then explored with large-scale MD simulations in the μs time scale, in which more than 260 000 CG segments were used to represent over 120 000 molecules. It was observed that only at these larger scales can one fully appreciate the effects of clustering and aggregation in the system. The results indicate that the propensity for asphaltene aggregation is significantly affected by the molecular morphology of asphaltenes.