ESPEYB25 14. The Year in Science and Medicine Genomics – Long-Read Sequencing (7 abstracts)
Am J Hum Genet. 2025;112(2):428-49 doi: 10.1016/j.ajhg.2025.01.002
Brief Summary: This study evaluated the added diagnostic value of long-read sequencing (LRS) in a rare disease context using scalable Oxford Nanopore LRS. It generated high-quality sequences (with ~36× coverage and 32 kb read N50 per sample) of 98 individuals from 41 families. The custom Napu bioinformatics pipeline assembled complete genomes, phased variants, and assessed methylation, detecting ~280 genes and 5 known Mendelian disease-associated genes inaccessible to short-read sequencing (SRS). With LRS it also detected complex structural variants, tandem repeat expansions, and epigenetic alterations. Notably, 87% of protein-coding genes were fully phased, allowing confident interpretation of compound heterozygosity, de novo events, and mosaicism. The study achieved a molecular diagnosis in 11 probands with diverse genetic etiologies.
The study demonstrates that LRS can serve as a comprehensive, single-test alternative to traditional multi-step genetic diagnostics for rare disease diagnosis, showing its ability to detect a wide range of variant types, including those inaccessible to SRS, while also enabling phasing and methylation analysis in a single assay. Although some methylation outliers and structural variants could not be confidently linked to disease due to limited functional annotation.
The open-source Napu pipeline and cost-effective protocol pave the way for broader clinical adoption of LRS. But, LRS still has lower base-level accuracy than SRS, particularly for small indels in homopolymer regions, which may lead to missed or miscalled variants. Interpretation of LRS-exclusive variants is hindered by the lack of long-read-specific population and clinical variant databases. Thus, larger, diverse population-scale LRS datasets are needed to improve variant interpretation and frequency estimation. Further development of somatic variant callers and tools for interpreting methylation and noncoding variation will enhance the diagnostic power of LRS in the future.