Long‑read sequencing improves structural detection

Long-read sequencing improves structural detection Cancer labs have spent the last decade getting very good at reading DNA in tiny fragments. That works well for spelling mistakes in the genome, but it is much less reliable for bigger problems like broken, flipped, duplicated, or relocated stretches of DNA. Those larger changes often decide a diagnosis, a prognosis, or whether a patient is eligible for a targeted drug. (nature.com) That is the opening for long-read sequencing. Instead of chopping the genome into very short snippets and trying to reconstruct the original from millions of puzzle pieces, long-read platforms read much longer stretches at once. In practice, that gives laboratories a better shot at seeing structural variants directly rather than inferring them from indirect clues. (nature.com) Structural variants are the category that includes deletions, insertions, inversions, translocations, and other rearrangements typically larger than 50 base pairs. Copy-number variants are a related class in which sections of DNA are gained or lost, changing how many copies of a genomic region a cell carries. In tumors, those events can activate oncogenes, disable tumor suppressors, or create gene fusions that point to specific therapies. (nature.com) Short-read sequencing struggles with these events for a simple reason: many of the hardest variants live in repetitive or difficult-to-map regions of the genome. A short read may land in several plausible places, so the software has to guess. A 2024 comparison in *Human Genome Variation* found that recall for structural-variant detection with short-read methods was significantly lower in repetitive regions, especially for small- to intermediate-sized events, than with long-read methods. (nature.com) PacBio has been one of the main companies pushing the clinical case for long reads, especially through its highly accurate “HiFi” reads. On PacBio’s own variant-detection materials, the company says HiFi sequencing supports high precision and recall for single-nucleotide variants, insertions and deletions, structural variants, and copy-number variants, including in difficult-to-map repetitive regions. Company claims should always be read alongside independent studies, but the direction of travel matches what the broader literature has been showing. (pacb.com) Independent research has been moving the same way. A 2025 *Nature* study using long-read sequencing in 1,019 people from the 1000 Genomes Project reported more than 100,000 sequence-resolved biallelic structural variants and 300,000 multiallelic variable-number tandem repeats, underscoring how much genomic variation becomes visible when reads are long enough to span complex regions. (nature.com) The clinical angle gets sharper in cancer, where genomes are often badly rearranged. A 2025 *Nature Methods* paper describing the SAVANA workflow focused specifically on somatic structural variants and copy-number aberrations in long-read data, noting that cancer genomes can contain patterns involving hundreds of structural variants across multiple chromosomes. That is exactly the sort of architecture that short reads can fragment into ambiguous signals. (nature.com) This is where molecular cytology enters the picture. Many patients with advanced cancer never have a large surgical specimen. Instead, the only material available may be a fine-needle aspirate, a fluid sample, a smear, or a cell block prepared from a minimally invasive procedure. Those samples are already central to diagnosis in lung cancer, thyroid disease, metastatic lesions, and other settings where speed and tissue conservation matter. (nature.com) Cytology has always been a “do a lot with a little” discipline. The College of American Pathologists notes that fine-needle aspiration is less invasive than a larger tissue biopsy and can provide enough material not only for morphology but also for ancillary studies including molecular testing. The same source emphasizes that rapid on-site evaluation can help direct extra passes specifically for molecular work, which is a practical advantage unique to cytology workflows. (cap.org) Recent clinical data show how far that field has already come. In a 2025 *Nature Communications* study from Memorial Sloan Kettering, 4,871 prospectively sequenced cytology samples were analyzed for comprehensive genetic profiling, and optimized workflows achieved success rates up to 93%, with overall performance similar to surgical samples for identifying clinically relevant genomic alterations. That is important because long-read methods do not arrive in a vacuum; they are landing in a cytology ecosystem that is already becoming molecular by default. (nature.com) The specific value of long reads for cytology is not that they replace every existing assay. It is that they expand what a lab can confidently offer when the question is dominated by structure rather than sequence spelling. If a case hinges on a large rearrangement, a complex fusion, a copy-number change, or a highly repetitive genomic region, long-read sequencing can turn a borderline or incomplete genotyping result into a resolved one. (pacb.com) That could change the analytic menu for difficult specimens. Today, many labs split the work across multiple methods: short-read next-generation sequencing for small variants, fluorescence in situ hybridization for some rearrangements, and separate copy-number approaches when needed. Long-read platforms raise the possibility of consolidating more of that into a single sequencing strategy, particularly for hard cases where standard testing leaves uncertainty. (frontiersin.org) There are still practical limits. PacBio’s whole-genome workflow materials describe starting with microgram-level amounts of unamplified genomic DNA and inserts in the 15 to 18 kilobase range, which is not trivial for every cytology specimen. Cytology samples are often scant, and cell blocks can carry processing artifacts or low-level contamination, so preanalytic handling remains a major determinant of success. (pacb.com) So the near-term story is not that every aspirate will suddenly go to long-read whole-genome sequencing. The more realistic shift is selective adoption: referral testing for structurally complex tumors, rescue testing when short-read panels are negative but suspicion remains high, and expanded use in cancers where fusions, amplifications, or rearrangements drive treatment decisions. (nature.com) For pathology laboratories, that means the menu may start to separate into two layers. Routine high-volume genotyping will still favor established short-read panels for many use cases, but a second tier of long-read testing is becoming easier to justify for the cases that break conventional workflows. In molecular cytology, where specimen quantity is limited and every assay has to earn its place, better structural detection is not just a technical upgrade. It is a different answer to the question of what can be learned from a very small sample. (cap.org)