I have found a small (but nefarious) error in the samples table which affects selection of samples in the app if samples with id>1472 are included. This apparently escaped previous testing. You can find the fix here (in the meantime I will upload an update in Mendeley data).