Understanding drug selectivity is crucial for designing effective medications that target specific proteins while minimizing off-target interactions, thus optimizing therapeutic outcomes. Although biochemical and cellular assays provide valuable insights on functional characterization, they offer limited understanding on molecular details of selectivity, particularly for targets within large families of structurally and functionally similar proteins. In such situations, carefully designed computational studies can complement experimental approaches, bridging gaps and deepening our understanding of drug selectivity mechanisms. In this study, Molecular Dynamics (MD) simulations along with structural bioinformatics methods have been used to examine the selectivity of Siponimod, an FDA-approved drug, towards sphingosine-1-phosphate receptor 1 (S1PR1) to treat Multiple Sclerosis (MS). Specifically, we investigated why Siponimod activates S1PR1 but not S1PR2, even though both receptors belong to the same sub-family of G Protein-Coupled Receptors (GPCRs). Contrary to the previous hypothesis based on molecular docking, MD simulations showed that Siponimod can bind to S1PR2 with anin silico-determined affinity comparable to that of S1PR1. Comparing the dynamics of Siponimod-bound S1PR2 with S1P-bound S1PR2 revealed that the transmission switches necessary for downstream biological activity are activated by S1P but not by Siponimod. Whereas, both S1P and Siponimod activate those transmission switches in S1PR1. A few crucial residues in S1PR2 were also identified that can be leveraged to optimize molecules to bind selectively to S1PR1 over S1PR2. Through our study, we showed thatin silicoapproaches can help in understanding the ligand induced structural changes and aid in developing and optimizing drugs for GPCRs, the most abundant class of membrane proteins in the human genome and the largest family of membrane receptors targeted by approved drugs.