The entire study is predicated on the claim that NIPT data, which is a mixture of ~90% maternal and ~10% fetal cell-free DNA, can be treated as a proxy for the maternal genome alone. The authors state: “the genetic information it contains predominantly reflects the maternal genome” (Page 3) and their entire “NIPT-human-genetics workflow” is built for “genome-wide association analysis of maternal genomes.”
This is a massive oversimplification and a critical error. By treating the mixed signal as purely maternal, the authors are systematically confounding maternal genetic effects with fetal genetic effects in every single analysis, from variant calling to GWAS.