When wanting to use on psychiatric PGS(s) to predict broad outcomes I usually use Schizophrenia, risk taking, or depression PGSs because all these GWAS are well powered. I was always wondering if it would be better to use p-factor (summing up all psychiatric liability gained by genomic-sem), and why? The thing is that I would end up with 4 factors/PGSs, which might be less well understand and defined in the field. Any advice?