Running Bases

DNA sequencing has come a long way in the 30 or so years since the development of the still widely used Sanger dideoxy method and the once-popular Maxam-Gilbert method. Many technological advances, including improvements in cloning, automation, and PCR have contributed to more rapid and less costly sequencing. The initial sequencing of the human genome was only the beginning—the genomes of multitudes of organisms of all kinds are being sequenced. Progress continues in the development of faster and cheaper methods in pursuit of a new goal, the “$1000 mammalian genome.”

Another Revolution

“Sequencing is really going through a revolution,” according to Edward M. Rubin, Director of the Genomics Division of the Lawrence Berkeley National Laboratory, Berkeley, CA, and Director of the Department of Energy's (DOE) Joint Genome Institute (JPI).

“Everything up to a year and a half ago was Sanger,” he says, noting that now non-Sanger sequencing has improved to the point that the JPI's Production Genomics Facility is mothballing many of its Sanger machines. He observes that the newer machines are good for sequencing 500–800-bp segments with greatly reduced cost and enormous increase in throughput compared with dideoxy sequencing. The technology works well now, Rubin says, for projects involving resequencing, where there is a reference sequence. “The challenge is de novo sequencing, where there is no reference. It's very complicated to assemble a genome that has not been previously sequenced. This will remain the realm of big sequencing centers.” The amount of data generated is still daunting, he says. A unique advantage to large sequencing centers is the ability to digest these data, whereas in the past, large centers had their advantage in having lots of robotics to handle the large volumes of liquids required for sequencing reactions. The independence of newer methods from requiring cloning of DNA in bacteria has eliminated that particular advantage.

Image 1.

Image 1. (Click to enlarge)

At JPI, the major sequencing focus is on organisms of interest to the DOE, including microorganisms and higher organisms (e.g., plants, that may be useful in the production of energy) for cellulose breakdown and fermentation of sugars, and for bioremediation. This distinguishes the JPI's activities from those of other large centers (e.g., the National Institutes of Health) where the emphasis remains on human biomedical applications.

Novel Technology

Nanopore technology is the basis of the novel methodology that Daniel Branton, Higgins Research Professor of Biology, Department of Molecular and Cellular Biology, Harvard University, Cambridge, MA, hopes will eventually be used for direct electronic sequencing of single DNA molecules. This method is based on detection of sequential bases as a DNA molecule translocates through solid state nanopores. One key concern is the detection method. Branton's group is investigating the use of transverse currents across the nanopores. Previously, they looked at how the DNA molecules blocked ionic currents through the pores. This was limited in the amount of current that was available for measurement at the speeds that would meet the high-throughput required for the $1000 mammalian genome.

The use of tunneling currents, in which DNA molecules are translocated through a tunneling gap, and electrons tunneling into the DNA as the molecules pass are detected, looks promising. This is analogous to a tunneling microscope, although instead of moving a probe past material, the material (DNA) is moved past the probe. The nanopores are made of carbon nanotubes, which are very stable and have interesting properties, Branton says, whereas metallic probes have substantial limitations. DNA binds naturally to carbon nanotubes, which he expects will allow orientation of the bases. Sequencing will depend on reproducible orientation of the DNA molecules between the emitter and receiver. “There's still quite a bit of basic research that needs to be done,” he concedes, noting that in the last year, they have shown that the bases do indeed orient in the nanotubes in a specific manner. They are now working out the binding energies required to release each base and to move from one base to the next along the DNA strand. “We have an overall picture of how it will work,” he says. “In principle, there is no limit to the length of a molecule that can be sequenced. This is a major advantage of this method, and it won't be as dependent on computer-based 'pasting' of sequences” as methods that generate short sequences.

Next Steps

Jay Shendure, Assistant Professor, Department of Genome Sciences, University of Washington, Seattle, WA, agrees with Rubin that the field of sequencing has changed very quickly in the last couple of years. “We've gone from a period where people were skeptical that anything would replace Sanger, to it being a matter of time,” he says. He was involved in the development of multiplex polony (polymerase colony) sequencing while in George Church's group at Harvard Medical School. In this method, colonies of short pieces of DNA are amplified using polymerase, which can be done on a support, such as beads or a slide or chip. The polonies are exposed simultaneously to bases labeled with different colors, and hybridization can be detected using a camera on a microscope as an optical scanner. Data are analyzed using software the group developed. Advantages of this method are that it can be performed in the average laboratory with standard equipment and reagents, and it is much less expensive than conventional sequencing.

Image 2.

Image 2. (Click to enlarge)

Shendure is starting a group to address some of the fundamental problems associated with sequencing. A few individual human genome sequences are not that informative. Sequencing needs to be done on a scale that allows for population studies. He suggests concentrating on interesting parts of the genome to the exclusion of other regions. For certain human diseases, for example cancer, he says, “The money is in the exons.” The problem is how to capture the subset of the genome of interest. Shendure's focus is on isolating large subsets of exons, for which he is developing a broadly useful targeted capture tool. He is collaborating with Mary Claire King, Professor of Genome Sciences and Medicine, Division of Medical Genetics at the University of Washington, to analyze genomic DNA samples from individuals with hereditary, early onset breast cancer. The presence of rare, highly penetrant alleles may explain the disease characteristics, but no obvious mutations have been detected. The pedigrees are too small to map, so their approach will be to sequence them all and “look for what pops out,” perhaps a rare exonic allele or frameshift.

Another application of inexpensive, multiplexed sequencing is to look at somatic mutations in the cancer genome, whether there are few samples, so statistical significance is difficult to assign, or for population studies of hundreds of individuals where it may be hard to determine the contribution of genetics. It may be necessary to distinguish the contribution of a few common variants from that of the collective contribution of a larger number of rare alleles.

The difficulty of having more data than one can keep up with is a fun problem to have, Shendure says. “There is pressure to build the right tools and to automate. The key is to do it right,” he says. “Sequencing data have the advantage of being very digital, unlike microarray data which are analog.”

Future Directions

Rubin predicts that technology will be developed to sequence much longer DNA molecules than is possible now. It may take a few years to market, but one goal is to be able to sequence a single DNA molecule. He thinks the $1000 genome is 10 years in the future, and that the quality of that sequence has to be taken into account. Shendure also believes that DNA sequences should routinely be given quality scores that reflect confidence in their accuracy.

Branton thinks it will take another three to four years to optimize nanopore sequencing technology. He predicts that ultimately a mammalian genome could be sequenced in 24 h on an array of 100 nanopores. “I'm going to retire after we solve this problem,” says Branton, who is an emeritus professor. “It will be the capstone of my career.”