AI-models have recently become a novel and popular toolkit in every data scientist's toolbox.  However, a current limitation of existing methods is that they are often very complex, and thus very difficult to interpret and trust.  In this research project we will develop novel "explainable" AI models that will lead to models that can be better interpreted, and thus also provide more information on the trustworthiness of a model.  As application domain we will focus on the recent research field of single-cell and spatial "omics".  These genomewide technologies allow measures at large scale molecular information of individual cells, as well as cells in tissues.  To gain more biological insights into these data, AI models are crucial, and in this context novel developments will allow building more interpretable models of cells and tissues that can lead to new functional biological insights.