Expanding the Genome in a Bottle Truth Set: Detection and Validation of Novel Low-frequency Variants Using High-accuracy NanoSeq
Zhang, Y.; Chao, H.; Niu, M.; Grochowski, C. M.; Kottapalli, K.; Bhamidipati, S. V.; muzny, d. m.; Gibbs, R. A.; Zong, C.; Doddapaneni, H. V.
Show abstract
HighlightsO_LINanoSeq-MBN achieves near-genome, Poisson-like coverage with minimal trinucleotide bias. C_LIO_LIExpands the GIAB truth set by up to 160k de novo variants. C_LIO_LIAdds a somatic layer to GIAB, enabling benchmarking and calibration of rare variants. C_LIO_LIHigh-CADD exonic and splice variants highlight value for surveillance and clinical triage. C_LI Somatic mutations record tissue molecular history and inform risk, prognosis, and therapy, yet their variant allele fractions often fall below the reliable detection limit of conventional short-read sequencing. In contrast, duplex sequencing technology featured by NanoSeq applies the principle of single molecule detection and thereby overcomes the limitation. However, the original NanoSeq protocol relies on the restriction enzyme-based genome fragmentation, which constrained its genome coverage to 30-40%. To enable whole-genome discovery with duplex-level fidelity, we pursued two complementary approaches to optimize the NanoSeq protocol: (i) a restriction-enzyme strategy densifies accessible sites using orthogonal 4-bp cutters; and (ii) a workflow using sonication followed by mung bean nuclease with T4 polynucleotide kinase, Klenow fragment and dATP/ddBTP mixture (NanoSeq-MBN) to blunt and repair/A-tailing DNA, while minimizing repair artifacts. We systematically benchmarked their performance using Genome in a Bottle (GIAB) gold-standard sample mixtures. As a result, NanoSeq-MBN achieved near genome-wide, Poissonlike coverage with minimal trinucleotide-context bias and ultra-high accuracy. Beyond variants already present in the GIAB truth set, NanoSeq-MBN identified approximately 120,000-160,000 de novo mutations per sample missing in the truth set, Notably, over 98% had orthogonal support in reanalyzed GIAB bulk Illumina HiSeq libraries. These novel variants extended GIAB from germline benchmarking to rare-variant discovery and calibration of subclonal detection. Functional annotation revealed enrichment of high Combined Annotation Dependent Depletion (CADD) scores mutations in exonic and splice-related regions. Variants intersecting ClinVar entries and OMIM genes highlighted potential for surveillance and clinical triage. Collectively, these results add a somatic layer to GIAB, enabling calibration of burdens and mutational signatures in lymphoblastoid lines and provide reference material for rare-variant assays. The NanoSeq-MBN workflow offers a path to whole-genome, high-fidelity discovery of ultra-rare somatic variation with relevance to clinical assay validation.
Matching journals
The top 5 journals account for 50% of the predicted probability mass.
Similar papers in this journal
- Cross-dataset pan-cancer detection: Correlating cell-free DNA fragment coverage with open chromatin sites across cell types 96%
- Comprehensive benchmarking of methods for mutation calling in circulating tumor DNA 96%
- Detection of isoforms and genomic alterations by high-throughput full-length single-cell RNA sequencing in ovarian cancer 96%
Similar papers in this journal
Similar papers in this journal
- A pan-cancer landscape of somatic substitutions in non-unique regions of the human genome 96%
- DeepMosaic: Control-independent mosaic single nucleotide variant detection using deep convolutional neural networks 96%
- Towards a Comprehensive Variation Benchmark for Challenging Medically-Relevant Autosomal Genes 96%
Similar papers in this journal
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.