In this project we will use the computational resources to develop, evaluate and use computational methods and protocols for drug design and discovery. The current main focus will be to investigate the use of machine learning methods to increase the throughput of virtual screening of ultra-large compound libraries with the aim to identify novel starting points for antimicrobial drugs, PET tracers and enzyme inhibitors. Furthermore, we will use molecular dynamics and free-energy perturbation to investigate ligand binding poses and support ligand design in our medicinal chemistry projects. In another subproject, DFT calculations will be used to investigate chemical reactions, with a focus on improving stereoselectivity in palladium-catalyzed reactions for the synthesis of potential drugs. The computational resources will be mainly used for a) ligand preparation b) docking of medium-sized compound libraries to obtain ML model training data, c) docking of large-size compound libraries for ML model evaluation and validation, d) MD simulations for investigating ligand binding poses and for ligand design using FEP, and e) DFT calculations for investigations of chemical reactions.