Donor-Specific Assemblies Reveal True Sequencing Error Profiles Across Seven Short-Read Platforms

Benchmarking nine short-read chemistries across seven DNA sequencers using donor-specific assemblies has established a robust method to isolate platform-intrinsic sequencing errors from authentic biological divergence. By evaluating two highly characterized benchmark samples, HG002 and COLO829BL, against sample-matched genomic references, researchers demonstrated that donor-specific assemblies accurately measure true instrument errors. This comparative strategy resolves a persistent obstacle in genomic research by directly separating authentic technological artifacts from natural inherited variation across different sequencing platforms and technologies.

However, the rapid expansion of commercially available sequencing platforms and specialized chemistries has created an increasingly urgent requirement to benchmark their performance under strictly standardized experimental conditions.

A major challenge identified by the study authors is that conventional reference-based evaluations routinely confound genuine platform sequencing errors with inherited genetic variation and reference bias. This confounding makes it difficult to isolate platform-intrinsic performance.

To eliminate these reference-induced artifacts, the investigators developed a comprehensive benchmarking framework utilizing donor-specific assemblies alongside two highly characterized benchmark samples, known as HG002 and COLO829BL. Across the seven DNA sequencers and nine short-read chemistries evaluated in the study, sequencing reads were measured against sample-matched genomic references derived specifically from each donor. This experimental design enabled the researchers to compare diverse sequencing technologies on an identical, sample-matched genomic foundation, ensuring that observed variations reflected authentic instrument behavior rather than underlying donor differences or sample divergence.

The benchmarking experiments revealed substantial differences among platforms across substitution, insertion-deletion (indel), read-position, and sequence-context error profiles. In substitution error evaluations, Element AVITI UltraQ and Roche SBX-D demonstrated the lowest substitution error rates among all evaluated sequencing systems.

In addition to overall error frequencies, the study uncovered pronounced platform-specific effects concentrated within low-complexity regions and specific trinucleotide sequence contexts. The authors observed prominent homopolymer-associated errors alongside context-dependent substitution skews that varied significantly depending on the sequencing chemistry employed.

The investigation further demonstrated that donor-specific references are essential for conducting unbiased base-quality recalibration, as they systematically minimize reference bias and establish reliable low-frequency variant-calling thresholds for cross-platform comparison. The authors concluded that donor-specific assembly-based benchmarking provides a robust and dependable framework for measuring true short-read sequencing errors across technologies. Ultimately, these results establish a comprehensive reference for the genomics community, providing empirical guidance for platform selection, experimental quality filtering, and the improved detection of rare somatic variation across genomic studies.

Subject of Research: Benchmarking short-read sequencing errors using donor-specific assemblies

Article Title: Short-read sequencing benchmarking with donor-specific assemblies

Article References: McGee, S. R., Smith, J. D., Frazar, C. D., Ryke, E., Vollger, M. R., Kwon, Y., Sohn, M.-H., Bennett, J. T., Eichler, E. E., Stergachis, A. B., & Wei, C.-L. (2026). Short-read sequencing benchmarking with donor-specific assemblies. Genome Biology. https://doi.org/10.1186/s13059-026-04331-4

Image Credits: AI Generated

DOI: 10.1186/s13059-026-04331-4

Keywords: short-read sequencing, donor-specific assemblies, genomic benchmarking, sequencing error profiles, DNA sequencers, base-quality recalibration, rare-variant detection, somatic variation, Short-read, sequencing, benchmarking, donor-specific