All products featured on WIRED are independently selected by our editors. However, we may receive compensation from retailers and/or from purchases of products through these links.
[Advances in DNA sequencing are crucial for the future of personal genomics, and approaches based on nanopores - tiny holes in a solid matrix, which can detect molecules passing through them - are a particularly promising area of innovation. You'll likely be hearing much more about nanopore-based sequencing over the course of 2011, and this guest post - from my Genomes Unzipped colleague Luke Jostins - provides the background you'll need to make sense of the announcements. –DM]
At last year's Advances in Genome Biology and Technology (AGBT) we saw a so-called "3rd generation" genome sequencing machines released by Pacific Biosciences and more information on a further machine from Life Technologies, often referred to as "Starlight", promised for, oh, about now.
These machines are "3rd generation" in the sense that they read a single strand of DNA at a time, as opposed to the clusters of DNA that the current second generation machines use, and can read much longer stretches of DNA much faster. However, these methods still use the same old-school optical technology that has been used since the first generation; they both identify the DNA sequence by shining a laser at DNA tagged with a fluorescent dye. This approach has drawbacks: it requires massive, expensive lasers, and tends to slowly fry the enzymes involved. The recently released Ion Torrent machine was one of the first to break this mould, by reading DNA using electrical signals given off by DNA synthesis reactions. However, Ion Torrent still does not read single molecules of DNA, is relatively slow, and can only read short bits of DNA.
There is an emerging technology that does not use optical detection but still reads single, long molecules of DNA at high speed. This technique is called Nanopore sequencing, and works by measuring electronic conductance across a membrane. No finished machines have been produced yet, but below the surface an army of researchers, in both public and private institutions, are working to make nanopore sequencing a reality.
Nanopore sequencing technologies work by feeding DNA through a small hole called a nanopore, embedded in a membrane. As the DNA moves through the nanopore, the conductance across the membrane changes: different base pairs change the conductance in different ways. You can think of this simply as there being a constant flow of electrons through the nanopore, and when DNA blocks the pore the flow of electrons decreases, changing the conductance. Different DNA bases are different sizes and shapes, and so the conductance changes a different amount. The current running through the nanopore is measured by a single electrode, with the eventual goal being to run many thousands of nanopores, each measured by its own single electode, on a semiconductor chip. As the nanopore approach is based on blocking the pore, it is pretty general; as well as reading DNA, it can also be used to identify when proteins have bound to a particular ligand, allowing you to measure protein expression.
The DNA can either be chopped up and spat into the pore by an enzyme (exonuclease sequencing), or instead is pulled through gradually (strand sequencing). The advantage of the former is that only one base is in the pore at a time, but the downside is that bases can get out of order. The exonuclease controls the reaction, feeding bases through one at a time, and ensures the bases go through at a manageble speed. In strand sequencing, the DNA strand is ratcheted, one base at a time, through the pore by a polymerase enzyme, though in solid-state sequencing the DNA may be ratched by a magnetic field.
The membrane in which the hold is formed can be either biological, such as a protein nanopore in a lipid membrane, or solid-state, such as graphene or silicon nitride with a hole in it. The pores themselves can also either be biological proteins, or solid-state pores. In general, solid-state membranes are thought to be more robust, can function in a range of environments, and can generally be linked up to and controlled easily by electronics; however, they are current difficult to fabricate in a consistent way. Making lots of identical proteins is easy, though proteins can be harder to control in real-time, and certain proteins can be very sensitive to the environment (though, interestingly, protein nanopores are basically indestructable). Hybrid systems, with a solid-state membranes with protein nanopores in them, are also being developed.
Unlike second generation sequencing, in which "clusters" of DNA are read (and get out-of-sync over time), each nanopore only reads one strand of DNA, and can, in theory, keep reading DNA for as long as it keeps being given it; in effect, the accuracy is independent of the read length. However, one challenge for nanopore sequencing is that the DNA may separate from the enzyme attached to the nanopore, similar to the way the enzymes of a PacBio machine stochastically die off from photodamage, meaning that read length is not unlimited. I've speculated in more detail about the technical advantages and limitations of the proposed 3rd generation sequencing technologies here.
Oxford Nanopore is one of the major players in the Nanopore sequencing field, and have their fingers in a number of nanoporous pies, working and collaborating with researchers on exonuclease, strand and solid-state sequencing. They recently produced a video explaining how both their exonuclease and strand sequencing will function. (The image on the top of the post is from this video).
More detail on both exonuclease and strand sequencing can be found on the Oxford Nanopore website.
GenomeWeb recently published a roundup of nanopore technology in 2010. It's pretty inspiring reading: over the last year, many obstacles on the road to a working machine have fallen. Researchers figured out how to use a polymerase to feed DNA through the nanopore, holding each base in place for a few dozen miliseconds, and then moving on to the next, and a group from Harvard demonstrated a proof-of-concept for solid-state sequencing using atom-thick graphene. New discoveries have led to potentially promising new nanopore enzymes, and perhaps most importantly, investment in research, from both public and private sources, has been stronger than ever.
The GenomeWeb article concludes:
Expert opinions on nanopore sequencing tend to be something like, "promising, but a long way from ready". Exactly how far from ready depends on the technology: companies working on exonuclease and strand sequencing may be a few years from producing commercial machines. For solid-state nanopores we are probably looking at 5 years or more.
As a warning, though, Oxford Nanopore has been working on its exonuclease technology for a long time; they demonstrated proof-of-principal nanopore sequencing in 2008. At around the same time, they signed an exclusive agreement with Illumina to distribute their machines based on their exonuclease method of sequencing, and have since given almost no information on exonuclease progress. Given how hush-hush Illumina managed to keep the development of the HiSeq platform, we really have no way of knowing how far from release the machine could be; it could be years away, or it could be announced before Christmas.
Luke Jostins is a UK-based postgraduate student working on the genetic basis of complex auto-immune diseases. He blogs at Genomes Unzipped and Genetic Inference.
