Sequencing of the human genome allowed for large-scale whole-genome population studies to identify the genetic underpinning of common diseases. While thousands of genome-wide association studies (GWAS) have now been published, a systematic validation of these efforts is lacking. Our goal in this study was to test if GWAS approaches do indeed identify genes/loci that modify disease severity. We focused on Parkinson’s disease (PD) a complex neurodegenerative disorder that severely impairs motor function with age. Meta analysis of GWAS data from ~25,000 PD patients and >100,000 age-matched controls identified single nucleotide polymorphisms (SNPs) of varying p-values (<10^-4) which corresponded to ~800 Drosophila melanogaster orthologs. Here we have systematically tested each candidate for PD-modifying effects by monitoring locomotor function and lifespan. Assessment of this dataset has allowed us to assess the validity of using GWAS to predict disease-modifying genes. We have been able to show that probability of a disease-modifying phenotype inversely correlates with GWAS p-value and the distance between gene and SNP. Thus we conclude that at least for large studies of heritable complex diseases, GWAS does indeed pinpoint disease-modifying loci and has allowed us to identify new conserved PD genes, which can inform on basic disease mechanisms or be targeted as novel therapies for this devastating illness.