We also compared the results given by four different analysis approaches: single-GWAS analysis with or without annotation, and two-GWAS joint analysis with or without annotation data. The Manhattan plots are shown in Figures 6 and 7. For single-GWAS analysis without annotation, GPA identified 13 SNPs and 391 SNPs with local false discovery rate for BPD and SCZ, respectively. By using the CNS set as annotation, GPA was able to identify 14 and 409 SNPs for BPD and SCZ, respectively, with the same fdr control. For joint analysis without annotation, the number of identified SNPs increased to 383 and 821 for BPD and SCZ, respectively. By using the CNS set as annotation, the number of identified SNPs further increased to 385 and 837 for BPD and SCZ, respectively. We investigated the BPD results in detail to evaluate the power of GPA in identification of functionally important SNPs. For single-GWAS analysis of BPD, GPA was able to identify SNPs located in the ANK3 gene. By using annotation data, the CACNA1C gene, which encodes an alpha-1 subunit of a voltage-dependent calcium channel, was
