Detecting and Quantitating Low Fraction DNA Variants with Low-Depth Sequencing

Abstract

DNA sequence variants with low allele frequencies below 1% are difficult to detect and quantitate by sequencing, due to the intrinsic error of sequencing-by-synthesis (NGS). Unique molecular identifier barcodes can in principle help NGS detect mutations down to 0.1% variant allele frequency (VAF), but require extremely high sequencing depths of over 25,000x, rendering high sensitivity mutation detection out of reach for most research and clinical samples. Here, we present the multiplex blocker displacement amplification (mBDA) method to selectively enrich DNA variants by an average of 300-fold in highly multiplexed NGS settings. On a 80-plex human single nucleotide polymorphism panel, mBDA achieves a 0.019% VAF limit of detection for single nucleotide variants, using only 250x sequencing depth, and detects human cell line contamination down to 0.07%. Using this technology, we constructed a 16-plex melanoma NGS panel covering 145 actionable mutations across 9 genes, and applied it to 19 fresh/frozen tumor biopsy tissue samples with high tumor fractions. We found low VAF mutations (0.2% to 5%) in 37% of the samples (7/19, 95% confidence interval 19%-58%). These results suggest that tumor heterogeneity could be significantly more pervasive than previously recognized, and can contribute significantly to acquired drug resistance to targeted therapies. We also validate mBDA panels on clinical cell-free DNA samples from lung cancer patients.Sequencing-by-synthesis, also known as next-generation sequencing (NGS), is one of the most powerful methods available today for highly multiplexed characterization of DNA at many genetic loci [1,2]. NGS is used routinely to detect DNA sequence variants with allele frequency of ≥ 5%. However, NGS struggles to report single nucleotide variants (SNVs) with variant allele frequencies (VAFs) of below 1% because all current NGS platforms have an average intrinsic error of at least 0.2% [3,4]. Error-correction methods based on unique molecular identifiers (UMIs) have been proposed and can circumvent the NGS intrinsic error limitation to achieve SNV limits of detection of about 0.1% [5-7], but these generally require sequencing to extremely high depths of 25,000x or higher. This causes the per-sample NGS cost of even moderate-size panels to be over $1,000 (e.g. the 70-gene Guardant 360 panel [8,9]), rendering sensitive detection of SNVs with VAF of below 1% unaffordable for many researchers, clinicians, and patients.Simultaneously, there is growing research and clinical demand for high sensitivity detection of low VAF DNA sequence variants. Examples of low VAF detection need include cellfree DNA profiling for non-invasive cancer therapy guidance and post-treatment monitoring [10,11], rare microbe detection for microbiome profiling [12,13], and persistent bacterial subpopulations contributing to antibiotic resistence [14,15] Digital PCR [16,17] and allele-specific PCR [18] can be used to detect and quantitate one or a few suspected DNA variants with VAF down to 0.1%, but cannot reasonably sc...

Oncogene Concatenated Enriched Amplicon Nanopore Sequencing for Rapid, Accurate, and Affordable Somatic Mutation Detection

“…Frequently, matched normal FFPE tissue samples will not be available, so using SAL alone for NS detection of low VAF so- [16,17]. The amplicons are subsequently appended with SAL adapters using PCR, and assembled into concatemers.…”

“…Frequently, matched normal FFPE tissue samples will not be available, so using SAL alone for NS detection of low VAF so-matic mutations is unlikely to be impactful clinically. The OCEANS method employs blocker displacement amplification (BDA) [16, 17] to allow more robust detection of low VAF somatic mutations without requiring a matched normal sample. In brief, BDA includes a wildtype-binding blocker oligonucleotide that competes with a PCR primer in hybridizing to DNA templates of interest.…”

“…Method and experimental results for the full OCEANS method. (a) Potential variants in multiple genetic loci of interest are first enriched using multiplex blocker displacement amplification (BDA) [16, 17]. The amplicons are subsequently appended with SAL adapters using PCR, and assembled into concatemers.…”

“…In this work, our bioinformatics pipelines make qualitative variant calls based on a combination of observed VRF and Clair score. In principle, the fold-enrichment could be calculated for each mutation in BDA [17], and the sample VAF could be inferred from VRF for OCEANS (Supplementary Section S4). In practice, however, the high NS error rates combined with the saturation of VRFs near 100% after BDA enrichment means that our quantitation dynamic range is relatively small.…”

“…1). The OCEANS method integrates the blocker displacement amplification (BDA) allele enrichment method [16,17] with a stochastic amplicon ligation (SAL), improving the VAF LoD by roughly 100-fold and throughput by roughly 10-fold over standard NS. The entire OCEANS method takes less than 10 hours from DNA to called variants, and the average sequencing cost per FFPE sample is roughly $3.75 for a 7-gene, 15 interest potentially bearing somatic mutations (red rectangle) are inefficiently and inaccurately sequenced using nanopore sequencing (NS).…”