The behavior of large, complex molecules in the presence of magnetic fields is experimentally challenging to measure and computationally intensive to predict. This work proposes a novel, mixed-methods approach for efficiently computing the principal magnetic susceptibilities and diamagnetic anisotropy of membrane proteins. The hierarchical primary (amino acid), secondary (α helical and β sheet), and tertiary (α helix and β barrel) structure of transmembrane proteins enables analysis of a complex molecule using discrete subunits of varying size and resolution. The proposed method converts the magnetic susceptibility tensor for all protein subunits to a unit coordinate system and sums them to build the magnetic susceptibility tensor for the membrane protein. Using this approach, we calculate the diamagnetic anisotropy for all transmembrane proteins of known structure and investigate the effect of different subunit resolutions on the resulting predictions of diamagnetic anisotropy. We demonstrate that amino acid residues with aromatic side groups exhibit higher diamagnetic anisotropies. On average, high percentages of aromatic amino acid subunits, a β barrel tertiary structure, and a small volume are correlated with high volumetric diamagnetic anisotropy. Finally, we demonstrate that accounting for the spatial position of the residues with respect to one another is critical to accurately computing the magnetic properties of the complex protein molecule.