DNA or RNA sequencing that reads long continuous stretches of nucleotides rather than short fragments.Long-read sequencing is a set of DNA and RNA sequencing methods that read much longer continuous stretches of nucleotides than the short-read technologies that dominated the field for years. A short read might cover a few hundred bases; a long read can span thousands to tens of thousands of bases in one pass. This matters because many features of a genome are simply too large to be reconstructed from small pieces: long repeated sequences, large structural rearrangements, and the full sequence of a gene from start to end.
With short reads, genomes are assembled by breaking the DNA into fragments, reading them, and computationally stitching overlaps back together. When the genome contains long repeats, the stitching becomes ambiguous and gaps remain. Long reads cross those repeats in a single read, so they anchor the assembly and resolve regions that were previously uncertain. In RNA, long reads can capture a whole transcript, showing which exons are joined together, rather than inferring transcript structure from fragments.
Long-read sequencing therefore explains and reveals things short reads often miss: full-length gene isoforms, large insertions or deletions, and the architecture of complex regions such as centromeres. It complements short-read data rather than replacing it, since short reads remain cheaper and highly accurate for many routine tasks.
Long-read sequencing became practical in the 2010s with the commercial introduction of single-molecule real-time sequencing by Pacific Biosciences in 2011 and nanopore sequencing by Oxford Nanopore Technologies in 2014. These platforms made it possible to read individual DNA molecules without breaking them into short fragments.
The main long-read platforms read DNA directly as it is synthesised or as it passes through a tiny pore. In one approach, a single DNA molecule is copied by a polymerase and the addition of each nucleotide is detected in real time from the light or electrical signal it produces. In another, the DNA strand is threaded through a protein pore and the change in electrical current as each base passes identifies the nucleotide. Because the signal is read from one intact molecule, read length is limited mainly by how long the molecule can be kept in the detector, not by fragmentation. The raw reads have a higher error rate than short reads, so they are usually corrected or used in many overlapping passes to produce an accurate consensus.
It is used to finish reference genomes, detect large structural variants in clinical genetics, and read full-length RNA transcripts in transcriptome studies.
The first complete, gap-free sequence of a human genome, announced in 2022, relied heavily on long reads to cover regions that short reads could not assemble.