Pre-templated instant partitions for sensitive sequencing: A novel approach for single-tube generation of multiplex, barcoded sequencing libraries

Detection of circulating tumor DNA (ctDNA) by barcoded next-generation sequencing (NGS) requires sophisticated library construction, and bioinformatics analysis to achieve ultra-sensitive (<0.1%) mutation detection. Preparation and sequencing of these libraries can be complex, time-consuming and expensive. Furthermore, DNA quantity is limited so efficient barcoding and amplification of the targets of interest is critical. In practice, many cancer detection and monitoring applications require longitudinal surveillance of multiple tumor or patient-specific mutations. For efficient sample utilization and cost-effectiveness, this will require customizable and flexible, multiplexed assay panels.

Here we present a novel NGS library preparation platform, PIPSenSeq (Pre-templated Instant Partitions for Sensitive Sequencing), that addresses these critical challenges. This approach takes advantage of molecular indexing for consensus-read sequencing and error reduction, and amplification in Poisson-distributed nanoscale partitions to reduce multiplex PCR primer conflicts and to enhance amplification uniformity across targets in multiplex PCR. Furthermore, PIPSenSeq library construction is performed in a fast, single-tube PCR format that eliminates sample loss steps, minimizes cross-contamination risk, and does not require expensive instrumentation or microfluidic consumables.

To demonstrate the efficacy of PiPSenSeq, we have developed and tested assays for >20 common tumor mutations. Each assay generates its anticipated, amplified library when tested in single-plex, and multiplexing assays consistently yields strong libraries with minimal off-target product formation. Sequencing of 12 - 20-plex libraries prepared from fragmented genomic DNA, with and without PIPs partitioning has been performed. As anticipated, PIPs partitioning significantly improved amplicon coverage uniformity (bulk CV=1.6, PIPs CV range 0.55-0.70) by eliminating competition between high and low efficiency targets. In addition, PIPs amplification reduces the likelihood of dropouts as compared to bulk. Multiplex PiPSenSeq libraries resulted in >90% on target alignment and >90% amplicon coverage uniformity. Using 25ng DNA standards, minor allele fractions of 0.1% were detected, with 94.4% sensitivity and 98% specificity.