Bipolar disorder (BD) includes a broad range of phenotypic features – most research examining its genetic underpinning focuses on the overall syndrome rather than its subphenotypic components. BD is a highly heritable disorder and while its genetic architecture is not fully elucidated there is increasingly strong support for a very polygenic component with partial polygenic overlap with schizophrenia (SZ). Here we present preliminary results from an analysis examining the association of SZ derived polygenic risk scores (PRS) across different BD subtypes (Bipolar Type 1(BPI), Bipolar Type 2 (BPII) and Schizoaffective disorder (SAD), and subphenotypes defined by the presence of psychosis and its associated level of mood incongruence.We used a UK sample of 3101 BP cases, collected using a consistently administrated interview protocol with OPCRIT confirmed lifetime diagnosis of Research Diagnostic Criteria (RDC - BPI, BPII or SAD). PRS were generated using 1000 genomes (phase3, 2014) imputed genetic data (INFO score>0.8, HWE>1e-6, MAF>0.01) trained on the results from Psychiatric Genetic Consortium (PGC2) SZ GWAS, pruned r2 <0.25, and computed across different SZ association p-value thresholds (range 0.0001 - 0.5). PRS were standardised and analysed with nominal logistic regression with RDC subtypes as dependant variables. To test whether psychosis mediated the association across subtypes, we repeated the analyses fitting psychosis together with PRS in the model and comparing it with the PRS only model. We also examined SZ-PRS association with psychosis in the whole sample and level of mood incongruent psychotic features in the subset of cases with psychosis (n =1352). All regression models were adjusted for age, sex and 1st 10 principal components derived from genetic data. Above analyses was run using DSM and ICDPRS with the SZ association p-value cut-off of 0.5, have shown significant association across the BD subtypes with relative risk ratios RRR = 1.28 (95%CI: 1.18-1.39, p<0.0001) and RRR = 1.18 (95%CI: 1.01–1.38, p=0.03) for BP1 and SAD, respectively, compared to BP2 as the baseline. Entering psychosis into the model attenuated the effect of PRS in the BP1 group but still remained significant (RRR = 1.19, 95%CI: 1.08-1.31, p<0.0001). In contrast the effect in the SAD group has become insignificant (RRR= 1.05, 95%CI: 0.89-1.24, p=0.6). In the whole sample of BD cases, PRS has shown association with presence of psychosis RRR=1.23 (95%CI: 1.15-1.34, p<0.0001). Examining cases with psychosis, we found PRS was associated with increasing level of mood incongruence (OR=1.15, 95%CI: 1.04-1.27, p=0.005). The results are ostensibly the same when PRS are generated using p-value thresholds from the range 0.05 - 0.5 , indicating these results are mostly driven by SNPs identified at p<0.05. Analyses using DSM and ICD show similar results.Preliminary analyses show a genetic - phenotypic association, between polygenic features of schizophrenia and BP characterised by the presence of psychosis, suggesting the phenotypic overlap between SZ and BD may in part be driven by polygenic overlap. This finding is further supported by the association of PRS and level of mood incongruence in the content of psychotic symptoms. These results suggest the PRS effect is not solely mediated though the psychosis subphenotype, at least in the BPI group. Our analyses show promise for the strategy of examining polygenic/phenotypic associations based on fine grained clinical characterisations