Inference of maternal allele inheritance via Bayesian hierarchical model in noninvasive prenatal diagnosis

Abstract

Noninvasive prenatal diagnosis (NIPD) poses a promising solution for detecting genetic alterations in fetus genome. However, the inference of the maternal allele inheritance in monogenic autosomal recessive disease is still challenging. Here the Bayesian hierarchical model is proposed to deduce the allele inheritance basing on haplotype frequency. The Bayesian approach, which does not depend on the knowledge of fetus DNA proportion in maternal plasma, provides accurate estimations on both real and simulated data; moreover, it is most robust than current methods in analyzing noisy or even erroneous data.