Molecularly imprinted polymers (MIPs) have potential as alternatives to antibodies in the diagnosis and treatment of disease. However, atomistic level knowledge of the prepolymerization process is limited that would facilitate rational design of more efficient MIPs. Accordingly, we have investigated using computation and experiment the protein-monomer binding interactions that may influence the desired specificity. Myoglobin was used as the target protein and five different acrylamide-based monomers were considered. Protein binding sites were predicted using SiteMap and binding free energies of monomers at each site were calculated using MM-GBSA. Statistical thermodynamic analysis and study of atomistic interactions facilitated rationalization of monomer performance in MIP rebinding studies (% rebind; imprinting factors). CD spectroscopy was used to determine monomer effects on myoglobin secondary structure, with all monomers except the smallest monomer (acrylamide) causing significant changes. A complex interplay between different protein-monomer binding effects and MIP efficacy was observed. Validation of hypotheses for key binding features was achieved by rational selection of two different comonomer MIP combinations that produced experimental results in agreement with predictions. The comonomer studies revealed that uniform, noncompetitive binding of monomers around a target protein is favorable. This study represents a step toward future rational in silico design of MIPs for proteins.