Abstract
Whole transcriptome RNA-sequencing is a powerful tool, but is costly and yields complex data sets that limit its utility in molecular diagnostic testing. A targeted quantitative RNA-sequencing method that is reproducible and reduces the number of sequencing reads required to measure transcripts over the full range of expression would be better suited to diagnostic testing. Toward this goal, we developed a competitive multiplex PCR-based amplicon sequencing library preparation method that a) targets only the sequences of interest and b) controls for inter-target variation in PCR amplification during library preparation by measuring each transcript native template relative to a known number of synthetic competitive template internal standard copies. To determine the utility of this method, we intentionally selected PCR conditions that would cause transcript amplification products (amplicons) to converge toward equimolar concentrations (normalization) during library preparation. We then tested whether this approach would enable accurate and reproducible quantification of each transcript across multiple library preparations, and at the same time reduce (through normalization) total sequencing reads required for quantification of transcript targets across a large range of expression. We demonstrate excellent reproducibility (R2 = 0.997) with 97% accuracy to detect 2-fold change using External RNA Controls Consortium (ERCC) reference materials; high inter-day, inter-site and inter-library concordance (R2 = 0.97–0.99) using FDA Sequencing Quality Control (SEQC) reference materials; and cross-platform concordance with both TaqMan qPCR (R2 = 0.96) and whole transcriptome RNA-sequencing following “traditional” library preparation using Illumina NGS kits (R2 = 0.94). Using this method, sequencing reads required to accurately quantify more than 100 targeted transcripts expressed over a 107-fold range was reduced more than 10,000-fold, from 2.3×109 to 1.4×105 sequencing reads. These studies demonstrate that the competitive multiplex-PCR amplicon library preparation method presented here provides the quality control, reproducibility, and reduced sequencing reads necessary for development and implementation of targeted quantitative RNA-sequencing biomarkers in molecular diagnostic testing.
Highlights
Next-generation sequencing (NGS) of RNA derived templates, or RNA-sequencing, is a powerful method with potential to rapidly advance discovery, development and implementation of transcript-based biomarkers in the clinical setting [1]
The percent variance explained (R2) was used as the objective criterion to compare the four methods across a range of the following quality control (QC) parameters corresponding to each transcript: 1) minimum number of sequence counts for an acceptable native target (NTi) or ISi measurement; and 2) the inter-replicate coefficient of variation (CV) across NCi
Among the endogenous gene targets, 82% (123 of 150) of designed assays for competitive amplicon library preparation produced one or more valid native target to internal standard (NT:IS) ratio measurements using ID 723 genomic DNA test material (Figure 1B and Table S1). For those assays with at least triplicate measurements, the ratio of native template (NT):IS sequencing reads decreased in direct proportion to increasing amounts of IS placed into the library preparation and at each titration point the inter-gene variation among NT:IS ratios was close to a 1:1 relationship (Figure 1B)
Summary
Next-generation sequencing (NGS) of RNA derived templates, or RNA-sequencing, is a powerful method with potential to rapidly advance discovery, development and implementation of transcript-based biomarkers in the clinical setting [1]. One way to avoid these effects is to routinely load less sample and/or use fewer amplification cycles for PCR based targeted sequencing approaches [8] This concept is similar to the earliest semi-quantitative end-point PCR methods, when PCR was performed on serial dilutions of target templates in an effort to find a concentration that yields detectable signal yet is not at plateau phase for any of the targets being measured [12,13,14]. Normalization between amplicons) from original representation is controlled because each target NT is measured relative to a known starting number of competitive IS copies (Figure 2 and Animation S1) Taking advantage of this characteristic, we designed amplification conditions that promote convergence of inter-transcript native amplicon representation toward equimolar proportions at plateau phase (Figure 2 with details in Methods section and Animation S1). The cost savings potential with this approach would be substantial because inter-gene variation in transcript representation commonly ranges more than one millionfold [2]
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
