USB-sized DNA sequencer is error-prone, but still useful

DNA passes through a pore, the basis of nanopore sequencing.

Nearly three years ago, a company called Oxford Nanopore made waves when it announced thedevelopment of DNA sequencing technology that was simple and compact enough to fit ona USB stick. The intervening years have been filled with product delays and apparent problems with read accuracy. This morning, however, a paper is being released that relies on the technology, and while the technology still has problems, it can help provide some medically relevant information.

Most methods of studying DNA perform what's called "sequencing by synthesis." They make a new copy of the DNA strand being sequenced, keeping track of which base is added to each location. This is highly effective, but it requires the use of consumables: each base added has to be supplied to the reaction, as does the enzyme that does the adding. All of this adds to the cost of the sequencing reaction, and it places a limit on how cheap we can make the process of sequencing a genome.

Oxford Nanopore's method is fundamentally different. Instead of adding new bases, the DNA strand to be sequenced is stuffed through a protein with a hole in the middle. As each base passes through this pore, its electrical properties are read, allowing the hardware to determine which base is going by. Although the pore will have a finite lifetime, the process doesn't need any specialized chemicals or enzymes and therefore has the potential to be very cheap.

It also has the potential to read the sequence of any length of DNA molecule. The most popular DNA sequencing technology, made by Illumina, is not commonlyused toread molecules that are much longer than 100 bases long, creating what's called "short read" sequences. It makes up for the short length by reading incredibly high volumes of DNA molecules.

But high volume doesn't entirely make up for things. In many areas of genomes, there are repeated sequences that are longer than 100 bases longthings like old, disabled viruses and transposable elements that hop around the genome. If a bit of sequence ends up in one of these repetitive elements, then there's no possible way to figure out where it resides in the genome. As a result, these repetitive sequences break up any genomes constructed solely from short reads, limiting our picture of the genome.

IfOxford Nanopore could get its system up and working, there would definitely be a use for it. But three years into things, and the new paper is still describing pre-release hardware. Gone is the claim of it being an actual USB device; instead, it's now referred to as "similar in size to a USB memory stick." But the biggest problem is the error rate; the system gets nearly a third of the bases it sequences wrong.

That may sound pretty useless, but it's not as bad as it seems. If both strands of the DNA molecule happen to be read, then accuracy gets bumped up to over 80 percent. And when combined with short-read Illumina data, it's possible to use it to help build more complete genomes.

The basic idea is that the long reads of the nanopore system can provide a scaffold for the entire genomeone that's error-filled but puts everything in the right order. The high-accuracy short reads from traditional sequencing methods can then fill in the scaffold with accurate data. The authors test this by sequencing a multidrug-resistant Salmonella strain. The drug resistance genes, in this case, were surrounded by repetitive DNA, leaving it unclear whether they were in the bacteria's genome at all, much less where they might be located.

In general, the approach of using both methods worked well. The number of gaps and unordered sequences went down, and a cluster of antibiotic resistance genes was identified in the regular chromosome of this strain. It wasn't entirely without problems, though. It turns out Illumina machines have problems when there are lots of G's and C's in a sequence, so there were a few areas that continued to be low quality. But the end result was a clearer picture of a medically relevant bacterial strain.

Here is the original post:
USB-sized DNA sequencer is error-prone, but still useful