An ideal targeted sequencing approach would allow the selection of any subset of DNA from a pool of molecules with minimal specialized library preparation. Current methods are sophisticated but not perfect, and can result in unnecessary sequencing of regions that are not of interest. Similarly, it is difficult to balance multiple sources of input DNA for a single sequencing library, such as are found in amplicon-based sequencing, and ratios cannot subsequently be modified after library preparation is complete. These problems cannot be detected until the sequencing run is complete and data are available for analysis.
Nanopore sequencing, auch as Oxford Nanopore's MinION, is based on the readout of electrical signals occurring when DNA nucleotides passing by protein nanopores. Change in the electric current is dependent on the shape, size and length of the DNA sequences. The DNA is moved through the pore by a motor protein. The small current changes, termed 'squiggles' are converted to DNA nucleotide information using base caller software, often located in the cloud. A team from the University of Nottingham now used signal processing techniques to map these squiggles to reference sequences, by passing this step.
The colleagues show that this squiggle matching technique can be performed at a rate that enables decisions to be made about the fragment of DNA that is being sequenced before it has completely passed through the nanopore. Depending on the sequence, individual nanopores within the MinION can then be instructed to continue sequencing or to eject the current DNA fragment and start sequencing another. They show that this real-time selective sequencing, or as some have called it DNA tasting, can reduce the time needed to sequence DNA fragments or enable the targeted analysis of fragments in mixed samples.
The anticipated speed increases in nanopore sequencing ('fast mode' is currently 300 b/s and may outpace this in the future) and scaling of the MinION to 3,000 channels and PromethION to 144,000 channels will challenge the implementation of Read Until and require algorithmic enhancements and computational power. However, we expect that selective sequencing will enable new approaches such as exon sequencing without target capture, controlled depth of coverage over entire genomes, counting applications for RNA-seq and many applications that have yet to be imagined.
I have one - DNA barcoding. Imagine a world in which you can identify any species on the spot, in an instant, anywhere on the planet. Sounds familiar? This was a slogan that stand at the beginning of iBOL only six years ago. We're getting close!