Checkpoint inhibition to treat cancer is now widespread across cancer types and stages, but response rates remain suboptimal, and immune related adverse events are a major challenge. I will try to provide a sythensis of current biomarkers of cancer immunotherapy efficacy and toxicity from a clinician and a mechanistic scientists’ perspective and highlight some vignette’s of our contributions viz. HLA-A*03 as a poor predictor of response (Naranbhai et al, Lancet Oncology 2022), IL-7 genetic variants as predictors of IRAEs (Groha et al., Nat medicine 2022) and expression variation in the immunoproteasome system as predictors of response (Ravi et al., Nat Genetics 2023). The focus will be on focussing efforts towards what specifically is needed, in my opinion, to propel the field forward in validating and leveraging these and other biomarkers to improve the utility of checkpoint inhibitors for patients.