ORs of the top and bottom PGS quintiles relative to ORs of the middle three quintiles in the MPN patient cohort (red and yellow) and the UKBB-MPN cohort (blue and gray) for PV. Data are presented as the OR (change in odds per increase of 1 s.d. in PGS; solid dots) with the 95% CI (error bars). Mutation-stratified groups are shown wherever these data were available. n is the sample size of PV cases in a specific group. Notably, PV negative for JAK2 mutation should be a rare diagnosis in PV (normally <5%), but several such individuals were present in the UKBB-WES cohort (n = 121 of 161; Supplementary Table 1). It may be that many of these individuals do not have an underlying clonal disorder such as PV but present with secondary erythrocytosis, due, for example, to smoking, alcohol or lung disease, where a germline predisposition to high HGB or HCT may contribute to blood count phenotypes that mimic PV (Supplementary Fig. 11). There was limited statistical power to demonstrate any significance for the PV subgroup positive for JAK2 mutation given the small sample size (UKBB-MPN cohort, n = 40). Furthermore, we cannot exclude the possibility that a mutant JAK2 clone was missed during sequencing in some of these individuals due to the low sequencing depth in UKBB or the timing of the blood sample relative to diagnosis.
