Droplet Digital PCR For Free Fetal DNA Analysis: Statistical Modelling And Evaluation For Non-Invasive Prenatal Diagnosis Of An X-linked Deletion.

Abstract

The discovery of cell-free fetal DNA (cffDNA) in maternal plasma has allowed the development of non-invasive prenatal diagnostic (NIPD). However, as cffDNA only represents a proportion (2-20%) of total cell-free DNA, it has been difficult to determine whether a fetus has inherited a maternal mutation using conventional techniques.
Single molecule counting techniques, such as digital PCR (dPCR), have been investigated for NIPD of such mutations as this requires precise quantification of mutant and wild type alleles and determination of whether there is an allelic imbalance. The direction and amplitude of the imbalance are directly influenced by the fetal genotype and the fraction of cffDNA present. Such a Relative Mutation Dosage (RMD) strategy has been used to infer fetal genotype from imbalances measured by dPCR in various clinical situations, although with a sensitivity which may not be suitable for clinical use. Existing studies have used first generation dPCR systems based on nanofluidic arrays, which provide a limited number of individual PCR chambers and may lack the precision and sensitivity required. In contrast, the more recent droplet digital PCR (ddPCR) approach appears promising in terms of number of partitions and cost but has not yet been investigated in the context of NIPD.
To address this, we adapted Power calculation methods based on existing statistical models of the variance associated with dPCR measurements and used them to predict the effect of number of partitions (chambers or droplets), total quantity of DNA available and fetal fraction, on the ability to discriminate between two fetal genotypes. Data will be presented that suggests: (i) that both number of partitions and sample quality are crucial determinants of the sensitivity achieved by dPCR in measuring allelic imbalances; (ii) that the larger number of partitions provided by ddPCR systems should in turn lead to a much improved precision and sensitivity.
Deletions and duplications within the X-linked dystrophin gene are a major cause of Duchenne and Becker Muscular Dystrophy (DMD/BMD). We therefore developed a dPCR assay design capable of measuring imbalances due to the presence or absence of a DMD deletion in a male fetus. We will show that determination of the male fetal fraction is achievable on two different ddPCR platforms (RainDrop and BioRad QX100), with a sensitivity and precision superior to that achieved on an array-based platform. Furthermore, a multiplexed assay was developed on the RainDrop platform that enables simultaneous determination of both male fraction and DMD balance (relative to X chromosome) in a single test. Optimisation, sensitivity and performance data of the DMD-specific assay using artificial mixtures of various genotype combinations relevant to clinical practice will also be presented and discussed in particular with respect to the number of droplets generated by each system and the predictions of our Power calculation tools.