Clin Chem. 2026 Jul 29:hvag074. doi: 10.1093/clinchem/hvag074. Online ahead of print.
ABSTRACT
BACKGROUND: Next-generation sequencing (NGS), also known as massively parallel sequencing, has become an essential tool across many areas of the life sciences, yet its application to low-frequency variant detection remains constrained by intrinsic error rates and the high cost of ultra-deep sequencing. To overcome these limitations, a range of technologies has been developed at both the library preparation and sequencing strategy levels to improve accuracy and boost sequencing efficiency, complemented by increasingly sophisticated computational pipelines for reliable mutation calling. Sequencing errors arise from multiple sources, including DNA damage, end-repair-associated misincorporations, PCR-derived errors, and base-calling inaccuracies.
CONTENT: Broadly, current highly accurate approaches fall into 2 conceptual categories: single-strand-consensus methods, which generate consensus from repeated observations of the same strand, and duplex-consensus methods, which require concordant evidence from both complementary strands of the original DNA molecule. This review summarizes the principles underlying these error correction strategies and highlights the methodological innovations that reduce errors at each step of the work flow.
SUMMARY: Highly accurate sequencing approaches have substantially improved the detection of low-frequency variants by reducing technical artifacts and enhancing signal-to-noise ratios.
PMID:42525853 | DOI:10.1093/clinchem/hvag074