Under high-resolution PRS in particular, multiple tests are performed and so the P-value of the best-fit PRS will be inflated. Therefore, we perform a permutation study utilizing the SCZ and MDD data described above, and estimate an adjusted significance threshold for the best-fit PRS of P = 0.004 (Supplementary Note S4). Prior to a more extensive study, we suggest a more conservative significance threshold of P = 0.001 if using the best-fit PRS for association testing in PRS studies.
