Behavioral screening defines three molecular Parkinsonism subgroups in Drosophila

Parkinsonism is defined by motor dysfunction, but the specific upstream molecular causes of these clinical symptoms can vary widely. We hypothesize that these causes converge onto a limited number of core cellular pathways. To investigate this, we created a new collection of 24 genetically very well-controlled animal models of familial forms of parkinsonism. Using unbiased behavioral screening and machine learning we identified three clusters of mutants that converge on (1) mitochondrial function; (2) retromer/vesicle trafficking; and (3) proteostasis/autophagy. Genes within each cluster have a similar genetic interaction profile and compounds that target specific molecular pathways ameliorate dopaminergic neuron dysfunction in a cluster-specific manner. This suggests that familial parkinsonism can be stratified into three broad functional groups and our findings pave the way for targeted biomarker discovery and drug development.