A testing process for Arborist would be providing a custom Makefile some subset of the inputs and then checking the output after running through the current codebase.
Data should include trick organisms/species and ones that include protein tree features like MHC nodes, fragments, and allergens.