I am using PRSice2 on the UKBB dataset. The phenotype is generated for a subset of patients, and it is absent for the remainder of the cohort. In total, there are over 400k absent phenotypes, so there are over 400k NaN values for the phenotype. However, when I run PRSice, it states that only 156 participants do not have phenotypes -- when in reality it should be over 400k. PRSice is able to correctly read the number of participants with phenotypes. Has anyone run into this problem or has a solution?