Treatment of inflammatory diseases is complicated by the fact that many different molecules and cell types are involved, and that these may differ at different times. Ideally, treatment would target an early molecule in an early cell type. This would require a multi-cellular disease model, showing in turn how all molecules in all cell types interact throughout the disease process. We recently described disease models in a single cell type, T cells, showing early and late disease processes in two inflammatory diseases, allergy and multiple sclerosis (Science Transl Medicine 2014 and 2015, Cell Reports 2016). The models were based on studying the activity of all human genes in T cells. We identified several molecules that could be suitable for targeted therapy. One of these was validated in an animal model of allergy, as well as by in vitro treatment of cells from allergic patients with a specific antibody. We now want to go further by constructing multicellular models of allergy and multiple sclerosis. The models are like puzzles where the pieces (=cells) are connected by linking molecules known to interact with each other. This will be done at different times to find out which cells start the diseases. We will search for 'starter molecules' and experimentally study whether these can be used for targeted treatment. If successful, the strategy may be transferable to find new medicines for all inflammatory diseases.