Nanopore sequencing can read stretches of DNA thousands of bases long. Such a capability is particularly useful for sequencing repetitive regions of the genome. But nanopore sequencing tends to be less accurate than other DNA sequencing methods at identifying individual bases. To get around this problem, researchers have previously taken a hybrid approach in which they use short reads from other technologies to correct long reads before assembling the sequence. Jared T. Simpson of the Ontario Institute for Cancer Research and the University of Toronto and Nicholas J. Loman and Joshua Quick of the University of Birmingham, in England, have for the first time assembled a bacterial genome using just nanopore sequencing data (Nat. Methods 2015, DOI: 10.1038/nmeth.3444). To improve accuracy, the researchers used only sequences for which both strands of a DNA segment had been read. The researchers lined up the sequence by detecting overlaps between reads and used an alignment process to correct base identification errors. After assembling the corrected reads into a contiguous sequence, the team used additional signal processing to improve accuracy of base identification to 99.5%.