Accurate quantification of placental (fetal) fraction by tissue specific cell-free DNA analysis.

Abstract

Background:Noninvasive liquid biopsy analytes such as cell-free DNA (cfDNA) have advanced the field of precision medicine. It is most widely used in noninvasive prenatal testing (NIPT) to screen for untreatable aneuploidies from maternal plasma. An important parameter for the success of the test is being able to accurately calculate the fraction of fetal cfDNA (FF). Recent advances in analysis of cfDNA’s fragmentation patterns have shown to reveal the tissue of origin which can be used to quantify the tissue’s fraction. Here we apply this concept to estimate the FF from shallow whole genome sequencing (sWGS). Method: In this study, we present an approach to quantify the placental fraction which is a surrogate for fetal fraction (FF). This is because the majority of fetal DNA in maternal plasma originates from placental tissue. We perform cfDNA fragmentation analysis using cfDNA tissue specific maps (cfDTM) to identify features that correlate with fetal fraction. We then apply machine learning strategies to accurately quantify FF from sWGS. The method is further validated using additional sample sets. Results: Fragmentation analysis of cfDNA using (cfDTMs) identified features that correlate highly with FF as calculated by a gold standard method – quantification of chromosome-Y derived fragments in mothers carrying male fetuses. We further identified Transcription factors associated with placental and hematopoietic origin. Finally, by training a model we were able to estimate the FF with very high accuracy. Furthermore, we show that the method is sex independent and is not affected by chromosomal aneuploidy. Conclusion: Using cfDNA samples we developed and validate a method of estimating the fraction of a tissue of interest using clinically routine sWGS techniques. Using placenta as our target tissue we were able to accurately quantify placental (fetal) fraction. This has immediate clinical application for NIPT to improve the accuracy of FF estimation reducing the number of false negatives. We anticipate this approach is generalizable which suggests that tissue-of-origin analysis of circulating cfDNA can be further applied to other clinically relevant use cases such as transplant rejection, tissue injury and in oncology as a prognostic and predictive biomarker.