Wenger, A.M. et al. Nat. Biotechnol. http://doi.org/dbgr (2019).

Long-read sequencing technologies have advantages in genome assembly, structural variant detection and haplotype phasing, but are less suited for single-nucleotide variant (SNV) and insertion/deletion (indel) calling due to the high error rate in comparison with short-read sequencing. Wenger et al., from Pacific Biosciences, optimized the circular consensus sequencing (CCS) protocol to achieve long, high-fidelity reads, in which they selected the SMRTbell library with fractions tightly distributed at 15 kb for high-coverage sequencing. The CCS library preparation was inspired and optimized based on their findings that polymerases have a better survival on damage-free DNA molecules. They employed the CCS protocol in sequencing the human genome (HG002) and achieved an average length of 13.5 kb, and an average accuracy of 99.8%. The CCS performance is comparable to, or better than, that of short-read sequencing in SNV and indel calling.