This thesis explores the computer-aided drug design of the M3-muscarinic acetylcholine receptor in three chapters. The first chapter focuses on molecular docking to identify novel agonists using the ZINC12 library. A model of the M3 receptor was built based on the M2 receptor. DOCK3.6 was applied using four docking strategies: (1) docking with the lead-like library (ZINC12) in the active conformation, (2) in the inactive conformation, (3) with the fragment-like library in the active conformation, and (4) using only quaternary ammonium ions. A receiver operating characteristic (ROC) curve with known ligands and decoys validated the docking results. Python scripts helped filter and analyze structures, revealing a library bias in DOCK3.6 with ZINC12. To test selected molecules, a Gq-activation assay from Cis-Bio was used. Since some molecules were weak binders, a competition assay was introduced, where a variable concentration of Carbachol was tested against a fixed high concentration of the test molecule. This study identified 53 novel ligands, including 25 agonists. The second chapter expands on these results. Due to the observed library bias, a collaboration with organic chemists allowed the synthesis of non-commercially available molecules. Molecular dynamics (MD) simulations showed that known ligands rarely formed hydrogen bonds with N6.52. To investigate this, acetylcholine was modified to tetramethylammonium, confirming its agonist activity. Nine subchapters explore different structure-activity relationships (SAR). These include modifications such as extending the n-alkyl chain, introducing branched or cyclic ammonium groups, and linking quaternary ammonium to a phenyl group. Additionally, one subchapter investigates the influence of other G proteins on Gq activation. A total of 108 molecules were tested in this thesis. The third chapter introduces a method for analyzing molecular dynamics simulations, called pose-population analysis. This tool quantifies the frequency of masked atoms at specific positions, allowing comparison between simulations. Two similarity functions were used: density similarity, measuring overlap at a position, and translational similarity, comparing spatial distributions. Six simulations were analyzed, focusing on acetylcholine and pilocarpine bound to M3 alone, M3 with Gq, and M3 with β-arrestin. The key residue, N6.52, showed low density similarity (10%) but high translational similarity (80%). As this residue is covalently bound to the receptor, its conformational space is restricted. The high translational similarity in- 2Summary dicates that different effector proteins stabilize distinct conformations. This tool enables quantitative comparisons of molecular dynamics simulations.