First let's briefly go over what the system is and how it works. The Nanopore MinION system is a small disposable sequencing platform that will sequence DNA by measuring changes in electrical conductivity as DNA is passed through a pore by a polymerase in the membrane (see this site for a helpful video explaining the process). Interestingly, the system IS NOT SEQUENCING INDIVIDUAL BASE PAIRS like most sequencing platforms, but rather is reading k-mers (small 'words' of DNA nucleotides, in this case 6 base pairs long; an example being 'GATACA'). This means that the MinION is picking up electric current signatures for 6 base pairs of DNA at a time, and assigning them based on complex probabilistic models for the most likely set of DNA that matches the signature. The system itself has 512 pores on the 1 square cm surface, with one molecule of DNA being processed through each pore at a time.
What about the performance itself? Well, the average length of the DNA sequences was 5kb and the longest reads were nearing 10kb, which is significantly below the company's goal of tens of kbs. The system also seemed to have systematic errors throughout sequencing, which is a headache for the bioinformaticians doing the analyses because systematic errors are more difficult to deal with than random errors. In the present data presentation, the MinION system was used to assemble two full bacterial genomes (E. coli and a bacteria from the Scardovia genus), but it was not able to make the assembly completely on its own. The assembly required sequences from Illumina based sequencing runs (these runs have shorter sequences but more of them). This lack independence is why the presentation felt a bit underwhelming, but I think this platform is still promising for two big reasons.
First, while we would all love to have this system worked out by now, it is still a work in progress and the error and length issues may be addressed. I got the feeling that Nanopore has the MinION platform itself worked out pretty well, but they are still working on optimizing the best ways to prepare the DNA for sequencing. The quality of the input DNA is vital and must be optimized for this specific system, and the results will surely improve as the DNA preparation protocols improve (as they say, "junk in, junk out"). In fact, Nanopore claims that the length of the DNA reads is limited by the fragmentation of the input DNA, and the lower sequence lengths in this first reported could be due to such fragmentation. Addressing this and the error related issues could make this a pretty efficient and ground-breaking system.
Second, I think the size of the machine is going to be a really import part of its overall success. This machine is small and feeds data directly to your laptop by plugging into a USB port. I can see this as being incredibly useful for scientists who may want to do some sequencing in the field, such as sequencing to see what samples might be important before collecting them all and taking them back to their lab. I can also see this as potentially having important implications in medical practices where diagnostic DNA sequencing could be performed in remote areas where large sequencing resources are unavailable. I think that the size of this platform is going to have interesting implications, and might open some new and exciting doors for field research and medical work, even if it only performs as good as its competitors as far as sequence length and abundance are concerned.
Erika Check Hayden (2014). Data from pocket-sized genome sequencer unveiled Nature DOI: 10.1038/nature.2014.14724
Twitter summary of the presentation
Oxford Nanopore Data and MinION: Valentines Day’s Gift to Genome Enthusiasts
The one and only Oxford Nanopore talk at AGBT 2014 – with real data
Picture 2 Source