Abstract: Functional Genomics, and CRISPR screening in particular, rapidly gained importance in the drug discovery process. Establishing a link of a gene to disease-relevant pathways in disease-relevant cell models is key for target identification and validation. CRISPR screening in primary and complex cell models, however, poses a number of challenges that must be overcome such as CRISPR reagent delivery, cell number limitation and endpoint complexity. We have developed a phenotypic screening platform to perform arrayed CRISPR screens in multiple primary and immune cell models to support projects across AstraZeneca’s portfolio. The platform allows us to combine different CRISPR reagent delivery methods with a wide-range of phenotypic arrayed assay endpoints, such as high content imaging or multi-colour flowcytometry. In this way we can tailor gene editing and the phenotypic assay to the required cell and disease model in order to deliver CRISPR screens and ultimately new drug targets in a variety of therapeutic settings. Embedding artificial intelligence (AI)/machine learning (ML) approaches has added considerable value to our phenotypic screening platform. Deep learning-based image analysis methods allow the separation of distinct phenotypes, such as “healthy” and “disease” in screens with high-content imaging endpoints. AI/ML can also be applied to mine knowledge graphs containing large disease-specific data sets to generate hypotheses that can then be validated using CRISPR. By performing CRISPR knock-out screens of knowledge graph-derived gene lists in disease-relevant models, we have been able to identify novel targets for chronic kidney disease (CKD) and idiopathic pulmonary fibrosis (IPF).