Antipsychotic-induced Parkinsonism (AIP) is an adverse effect that significantly limits the use of first generation antipsychotic drugs (FGAs). Frequency estimates vary from 10-40%. Second generation antipsychotics (SGAs) are associated with lower AIP rates but carry a significant risk of metabolic adverse effects. Thus, personalized treatment with antipsychotics is a key objective in pharmacogenomics. Several studies have sought genetic biomarkers for AIP. Most were retrospective and employed candidate variants in case-control association studies. Some associations have been reported but replication has not been consistent and a level of predictive power appropriate for clinical application has not been demonstrated. A case in point is the association of the functional rs4606 polymorphism in the RGS2 gene with susceptibility to early onset of AIP. This was first reported by our group in a prospective study of Israeli patients with schizophrenia (Greenbaum et al, 2007). The association was replicated in a retrospectively evaluated U.S. sample (Greenbaum et al, 2009) but not by other authors. We also obtained supporting evidence from a transgenic mouse model (Greenbaum et al, 2012a). Subsequently, we presented evidence for association of the intronic SNP, rs12678719, in the ZFPM2 gene with susceptibility to AIP. This was observed in a secondary analysis of CATIE data (Alkelai et al, 2009) and replicated in an independent U.S. case control sample (Greenbaum et al 2012b). Considering the two genetic variants together yielded a positive predictive value of 0.71 and a negative predictive value of 0.68 (odds ratio of 5.13) (Lerer et al, unpublished). To validate these findings we recruited a new sample of 404 Israeli schizophrenia patients on a stable regimen of antipsychotic treatment, of whom 383 had valid phenotypic and genetic data and were included in the data analysis (AIP positive: Simpson Angus Scale (SAS) score>0.3, N=206; AIP resistant, SAS ≤0.13, N=177). Recruitment was constrained to insure equivalent proportions of patients on FGAs and SGAs. Taking into account confounding variables such as age, gender, age at first antipsychotic treatment, ethnicity, antipsychotic drug dose in chlorpromazine units and treatment with anticholinergic medication, the results of this study showed no significant contribution of rs4606 in RGS2 and rs12678719 in ZFPM2 to AIP susceptibility. This was also the case when examining a more extreme phenotype. These findings illustrate the difficulty in establishing the clinical utility of common genetic variants as biomarkers for susceptibility to antipsychotic drug effects even with a carefully recruited, well-powered sample. They emphasize the need for large, prospective clinical studies with built in replication elements. A potential role of multiple, rare genetic variants should also be considered. In addition, epigenetic mechanism may contribute to inter-individual variability in susceptibility to AIP.