Determination of fetal DNA fraction from the plasma of pregnant women using sequence read counts

Abstract

This study introduces a novel method, referred to as SeqFF, for estimating the fetal DNA fraction in the plasma of pregnant women and to infer the underlying mechanism that allows for such statistical modeling.Autosomal regional read counts from whole-genome massively parallel single-end sequencing of circulating cell-free DNA (ccfDNA) from the plasma of 25 312 pregnant women were used to train a multivariate model. The pretrained model was then applied to 505 pregnant samples to assess the performance of SeqFF against known methodologies for fetal DNA fraction calculations.Pearson's correlation between chromosome Y and SeqFF for pregnancies with male fetuses from two independent cohorts ranged from 0.932 to 0.938. Comparison between a single-nucleotide polymorphism-based approach and SeqFF yielded a Pearson's correlation of 0.921. Paired-end sequencing suggests that shorter ccfDNA, that is, less than 150 bp in length, is nonuniformly distributed across the genome. Regions exhibiting an increased proportion of short ccfDNA, which are more likely of fetal origin, tend to provide more information in the SeqFF calculations.SeqFF is a robust and direct method to determine fetal DNA fraction. Furthermore, the method is applicable to both male and female pregnancies and can greatly improve the accuracy of noninvasive prenatal testing for fetal copy number variation. © 2015 John Wiley & Sons, Ltd.