Comparison of DNA Quantification Methods for Next Generation Sequencing

Next Generation Sequencing (NGS, also called High Throughput Sequencing) has facilitated a surge of new discoveries, from genotyping cohorts of patients for the identification of disease-associated mutations to the exploration of cellular processes and chromatin landscapes1,2,3. Whole-genome sequencing4,5,6,7,8 has become the method of choice for acquiring large data sets for broad genomic analyses, including the International Network of Cancer Genome Projects9, Cancer Genome Atlas Research Network10, the Integrated Genomic Analyses of Ovarian Carcinoma11, and the Encyclopedia of DNA elements project (ENCODE)12. With such large data sets and discoveries hinging on this technique, it is imperative that each step of NGS is optimized, accurate, and cost-effective. Current methods for quantifying DNA library preparations for NGS use a variety of techniques (Table 1): UV absorption (e.g., Nanodrop); intercalating dyes (e.g., QuBit; Invitrogen, SYBR Green); 5′ hydrolysis probes (e.g., Taqman®) coupled with real-time quantitative PCR (qPCR; Kapa Biosystem) or droplet digital emulsion PCRs (ddPCR; Bio-Rad). For most of these methods, large amounts of DNA are required to provide accurate titration. The primary reason why sequencing platforms require micrograms of library preparations (Illumina recommendation) is to ensure that enough material is available for quantification13,14,15. The amount of DNA needed for sequencing (12 pM per lane on average) is about 1000-fold less than what is needed for quantification. For certain libraries, this amount (i.e., microgram) of DNA is not available without excessive PCR amplification. Table 1 Titration methods used for Next Generation Sequencing. The generation of micrograms of DNA by PCR is a sensitive process subject to many variations. Increasing the number of PCR cycles leads to loss of sequence heterogeneity, potentially resulting in variable sequencing quality2,13,15; distortion of the original ratios/proportions of sequences and finally causing the loss of rare input molecules16. Also, sequence heterogeneity might be modified by PCR amplification altered by over amplification of shorter fragments, GC:AT content, and DNA/Taq polymerase bias17. Hence, most would agree that unnecessary amplification for the sole purpose of NGS library quantification should be avoided and cycle number should be minimized to prevent biases18. Our ability to reduce the number of cycles of PCR is hampered by the fact that current NGS library quantification methods are inadequate. To improve NGS library quantification, several methods have been developed. White et al. developed an original technique which utilizes 5′ hydrolysis probes and digital emulsion PCR using a chip to create ~1000 individual PCR reactions per chip19. In brief, they incorporated the complementary sequence of a 5′ hydrolysis probe to the forward primer sequence such that a hydrolysis probe can be used to detect library PCR products following end-point PCR (Fig. 1). This strategy and recent advances in emulsion PCR instruments allow one to combine accuracy, sensitivity, and scalability in library quantification. Droplet digital PCR devices, such as the QX100® ddPCR System (Bio-Rad Laboratories) can generate ~20,000 droplets and help ensure accurate quantification of DNA molecules from library preparations. Droplet digital PCR (ddPCR) generates thousands of droplets containing template DNA, partitioned and amplified within each individual droplet, allowing for smaller sample reaction sizes and also preserving rare molecules. After PCR, droplets are counted with a droplet reader and scored as either fluorescent or not, a binary readout is produced for the 20,000 droplets (hence the term digital). The ratio of PCR positive to negative droplets is calculated, and Poisson statistics corrects for the possibility of having one or more template molecules per droplet. The resulting quantification of input DNA molecules is given as a concentration [molecules of target per microliter of input PCR], and this concentration is easily converted to total number of input molecules of a given sample. DdPCR simplifies and improves the process of determining the number of input DNA molecules in a given sample compared to methods using standard curves and qPCR. Figure 1 Overview of the ddPCR-Tail strategy using UPLs® (a 5′probe-tailed primer). In this work, we combined the power of 5′ hydrolysis with ddPCR to achieve quantification of NGS libraries, and compared this method with samples quantified by other traditional methods. We describe the adaptation of the universal template probe digital PCR method to the BioRad QX100® ddPCR platform for the quantification of sensitive low abundance NGS libraries. We extend our report with the adaptation of the UPL technology (Universal Probe Library; Roche) to the same platform and provide a comparison between the current NGS titration methods. The methodology described here provides sensitivity and stability not achieved by other techniques. By directly comparing identical samples through a full sequencing run as a proof of principal pilot series of experiments, we present an end-point analysis that was missing regarding NGS quantification strategies.