Dear editor, Please find our manuscript entitled ”Knowledge graphs and explainable AI for drug repurposing on rare diseases” for your consideration. A key open challenge in ML-based drug-disease prediction is how to provide a human understandable explanation that can aid biologists in the generation of testable hypotheses in the lab. We developed rd-explainer a novel method that utilises knowledge graphs in combination with cutting-edge graph ML and XAI tools to provide semantic graphs as explanations supporting predictions. Graph neural networks is one of the most used algorithms in drug repurposing, but how to combine them with background knowledge and XAI tools for better interpretability is barely explored specially for the underrepresented group of rare diseases. We de- veloped a novel interpretable ML algorithm that allows graph neural networks to provide semantic explanations that resembles to human reasoning, and combine this neuro-symbolic method with disease specific knowledge graphs. Our approach is generic and can be applied in different rare diseases and can be enhanced by disease specific background knowledge. Using several evalua- tion tests and specific use cases, we demonstrate that our method can substantially improve the performance of drug-phenotype prediction. We believe that rd-explainer, as well as the underlying method combining knowledge represen- tation and graph-based ML and XAI, will have a broad impact in ML-based biomedical discovery, both in the specific application of drug repurposing prediction and in related areas such as rare disease research. Therefore, we believe our work is highly suitable for Neurosymbolic Artificial Intelligence. Please do not hesitate to contact us should you require any further information. With kind regards, The authors.