@article{hadi_zareie_atomic-resolution_2003,
	title = {Atomic-resolution {STM} structure of {DNA} and localization of the retinoic acid binding site},
	volume = {303},
	issn = {{0006-291X}},
	url = {http://www.sciencedirect.com/science/article/B6WBK-482YX0V-5/2/28cc5d68d897bece640fda09b341a9d1},
	doi = {{10.1016/S0006-291X(03)00298-5}},
	abstract = {
Single-molecule imaging by scanning tunnelling microscopy {(STM)} yields the atomic-resolution (0.6 Å) structure of individual B-type {DNA} molecules. The strong correlation between these {STM} structures and those predicted from the known base sequence indicates that sequencing of single {DNA} molecules using {STM} may be feasible. There is excellent agreement between the {STM} and X-ray structures, but subtle differences exist due to radial distortions. We show that the interactions of other molecules with {DNA,} their binding configurations, and the structure of these complexes can be studied at the single-molecule level. The anti-cancer drug retinoic acid {(RA)} binds selectively to the minor groove of {DNA} with up to 6 {RA} molecules per {DNA} turn and with the plane of the {RA} molecule approximately parallel to the {DNA} symmetry axis. Similar studies for other drug molecules will be valuable in the a priori evaluation of the effectiveness of anti-cancer drugs.},
	number = {1},
	journal = {Biochemical and Biophysical Research Communications},
	author = {M Hadi Zareie and Philip B Lukins},
	month = mar,
	year = {2003},
	keywords = {{DNA} {imaging,Drug} {binding,Retinoic} {acid,Scanning} tunnelling microscopy},
	pages = {153--159}
},

@article{hansma_atomic_2004,
	title = {Atomic force microscopy imaging and pulling of nucleic acids},
	volume = {14},
	issn = {{0959-440X}},
	url = {http://www.sciencedirect.com/science/article/B6VS6-4CDS7DK-4/2/b48944ee3d8a5e8a4d731dc885be8f73},
	doi = {10.1016/j.sbi.2004.05.005},
	abstract = {
Recent advances in atomic force microscopy {(AFM)} imaging of nucleic acids include the visualization of {DNA} and {RNA} incorporated into devices and patterns, and into structures based on their sequences or sequence recognition. {AFM} imaging of nuclear structures has contributed to advances in telomere research and to our understanding of nucleosome formation. Highlights of force spectroscopy or pulling of nucleic acids include the use of {DNA} as a programmable force sensor, and the analysis of {RNA} flexibility and drug binding to {DNA.}},
	number = {3},
	journal = {Current Opinion in Structural Biology},
	author = {Helen G Hansma and Kenichi Kasuya and Emin Oroudjev},
	month = jun,
	year = {2004},
	pages = {380--385}
},

@article{lindsay_potentiostatic_1992,
	title = {Potentiostatic deposition of {DNA} for scanning probe microscopy},
	volume = {61},
	url = {http://www.sciencedirect.com/science?_ob=ArticleURL&_udi=B94RW-4V8X9NB-D&_user=108429&_rdoc=1&_fmt=&_orig=search&_sort=d&view=c&_acct=C000059713&_version=1&_urlVersion=0&_userid=108429&md5=57dfed35023dfbf00aa2b86ed71a6ec7},
	doi = {{10.1016/S0006-3495(92)81961-6}},
	abstract = {{aDepartment} of Physics, Arizona State University Tempe 85287–1504. We describe a procedure for reversible adsorption of {DNA} onto a gold electrode maintained under potential control. The adsorbate can be imaged by scanning probe microscopy in situ. Quantitative control of a molecular adsorbate for microscopy is now possible. We found a potential window (between 0 and 180 {mV} versus a silver wire quasi reference) over which a gold (111) surface under phosphate buffer is positively charged, but is not covered with a dense adsorbate. When {DNA} is present in these conditions, molecules adsorb onto the electrode and remain stable under repeated scanning with a scanning tunneling microscope {(STM).} They become removed when the surface is brought to a negative charge. When operated at tunnel currents below approximately 0.4 {nA,} the {STM} yields a resolution of approximately 1 nm, which is better than can be obtained with atomic force microscopy {(AFM)} at present. We illustrate this procedure by imaging a series of {DNA} molecules made by ligating a 21 base-pair oligonucleotide. We observed the expected series of fragment lengths but small fragments are adsorbed preferentially. },
	number = {6},
	journal = {Biophysical Journal},
	author = {{S.M.} Lindsay and {N.J.} Tao and {J.A.} {DeRose} and {P.I.} Oden and Lyubchenko {YuL} and {R.E.} Harrington and L. Shlyakhtenko},
	month = jun,
	year = {1992},
	pages = {1570--1584}
},

@article{ryan_toward_2007,
	title = {Toward nanoscale genome sequencing},
	volume = {25},
	url = {http://www.sciencedirect.com/science?_ob=ArticleURL&_udi=B6TCW-4P77G02-1&_user=108429&_rdoc=1&_fmt=&_orig=search&_sort=d&view=c&_acct=C000059713&_version=1&_urlVersion=0&_userid=108429&md5=0db97e1437e06015aa259e07225ce06f},
	doi = {10.1016/j.tibtech.2007.07.001},
	abstract = {{aDepartment} of Electrical Engineering, University of Washington Paul Allen Center, {AE100R} Box 352500 Seattle, {WA} 98195, {USA} This article reports on the state-of-the-art technologies that sequence {DNA} using miniaturized devices. The article considers the miniaturization of existing technologies for sequencing {DNA} and the opportunities for cost reduction that ‘on-chip’ devices can deliver. The ability to construct nano-scale structures and perform measurements using novel nano-scale effects has provided new opportunities to identify nucleotides directly using physical, and not chemical, methods. The challenges that these technologies need to overcome to provide a {US\$1000-genome} sequencing technology are also presented.   The state-of-the-art technology in human genome sequencing uses Sanger sequencing {(Box} 1) and is capable of sequencing 89\% of the {DNA} in a typical run; the length of an individual read is 805 Q20 bases (sequencing bases with an error rate {\textless} 1\%) [1]. A high-throughput genome center typically performs 106 reactions every month; this throughput and corresponding read length approximates a mammalian-sized genome. To ensure accuracy and contiguity in the sequencing method, redundancy equivalent to six times the read length of the genome is required. The total expenditure for sequencing a typical mammalian-sized genome is six months in time and costs approximately {US\$12} million. Box 1. Sanger sequencing This method of sequencing exploits the different reactivity of deoxynucleotides and dideoxynucleotides. Deoxynucleotides {(dNTPs;} N represents the base A, G, T, or C) contain a reactive {3′-OH} functional group that reacts with the next nucleotide in the {DNA} polymerization reaction; dideoxynucleotides {(ddNTPs),} however, do not contain the same {3′-OH} functional group and terminate the polymerization when incorporated into {DNA.} By way of illustration consider a mixture of {dNTP} and {ddATP.} Polymerization of {DNA} that is complementary to the template {DNA} will continue until a molecule of {ddATP} is incorporated into the sequence. The product of this reaction is a distribution of lengths of {DNA} that terminate with an A nucleotide. The distribution of different {DNA} molecules is separated using capillary electrophoresis in a poly(acrylamide) gel, and a laser scans the capillary to excite a fluorescent marker (which is located on the primer that initiates synthesis or on {ddATP)} and is used to read the position of all A nucleotides in the sequence. This process is repeated for G, T, and C (using {ddGTP,} {ddTTP} and {ddCTP,} respectively), and the collective information is used to produce the sequence of the {DNA} template. This review describes recent trends in scaling traditional genome sequencing technologies to the microscale, as well as the introduction of nanoscale methods that could be used to sequence genomes. This review does not detail the significant advances in recombinant {DNA} engineering [2], fluorescent dye development [3], capillary electrophoresis [4], automation [5] and [6] and computation methods [7] that provide the infrastructure to quantify and interpret rapidly the data produced from a sequencing reaction. The method for sequencing {DNA} developed by Sanger et al. as well as its variations were crucial in obtaining the complete map of a human genome [8] and [9]. Recent advances in Sanger sequencing include the development of alternative methods to excite fluorescent dyes (e.g. pulsed multi-line excitation [10]) and new strategies for fluorescence detection (e.g. using the lifetime of fluorescence as the ‘signal’ from the dye [11]). Microfabricated devices that perform Sanger sequencing might reduce the cost of {DNA} sequencing (per read) [12], however, it remains unclear whether they can compete with Massively Parallel Strategies already in use. The most recent demonstration of on-chip sequencing does indicate significant reduction in the time to separate {DNA} (conventional capillary array electrophoresis takes hours, microelectrophoresis takes minutes, Figure 1) [13], however, the serial processing of samples into 96 or 384 microchannels will not compete with the parallel processing of 1.6 million wells in 454 Corporations's picotiter plates {(PTPs).} Clonal amplification is a key enabling strategy for microscale genome sequencing, which is used in both pyrosequencing {(Box} 2) and fluorescent in situ sequencing {(FISSEQ)} {(Box} 3). Clonal amplification exploits the size-confinement of micro-fabricated wells or a polymer gel; 454 Corporation uses {PTPs} to perform 105 {PCR} reactions simultaneously. {FISSEQ} uses polony [14] [e.g. the amplification of individual {DNA} molecules that are restricted in space in a poly(acrylamide) gel] to produce large numbers of identical {DNA} templates. The {FISSEQ} technology can be extended to include four {dNTPs} that contain unique cleavable fluorophores. This technology was shown to achieve a read length of up to 12 bases [15]. Another approach called {BEAM} [16] (beads, emulsion, amplification, magnetic) produces identical copies of {DNA} in an oil-in-water emulsion using {PCR} amplification of {DNA} attached to magnetic beads. Clonal amplification is a useful strategy for sequencing in general and, more recently, for patient genotyping and haplotyping [17], [18] and [19]. Box 2. Pyrosequencing Pyrosequencing is a method for sequencing {DNA} that converts pyrophosphate (a product from the reaction between a sequence of {DNA} and the next {dNTP} to be added in the sequence) to a photon via a chemiluminescent reaction [33] and [34]. The photon is detected in real-time from the {DNA} that is polymerizing along a template in response to the sequential addition of {dNTPs.} Box 3. Fluorescence in situ-sequencing Fluorescence in situ sequencing {(FISSEQ)} uses {dNTPs} that contain cleavable fluorescent dyes to sequence {DNA} [14]. When a fluorescent {dNTP} is incorporated into the {DNA} sequence the fluorophore is cleaved chemically or photochemically; the fluorescence from the cleaved fluorophore is used to interpret the sequence of {DNA;} the next fluorescent {dNTP} is incorporated and the process repeats itself. Sequencing by hybridization {(SBH)} uses oligonucleotide probes in a microarray format that exhibit different affinities for a target sequence of {DNA} [20]. {SBH} exploits microarrays to sequence {\textgreater} 109 bases with a read length defined by the length of the oligonucleotide probe. Improved hybridization of {DNA} to the surface-bound oligonucleotide probes can be achieved using dendritic molecules that attach the probes to the surface [21] and [22]. The dendritic molecules increase the spacing between individual probes, which results in increased hybridization between the probe and the target {DNA.} Optical trapping of two polystyrene beads, where one bead is attached to an {RNA} polymerase and the other bead is attached to a short strand of template {DNA,} can be used to sequence {DNA} [23]. This single-molecule method can detect the processive motion of the polymerase as it incorporates nucleotides along the template {DNA.} When the concentration of one type of nucleotides is limited, the polymerase stops on the template {DNA} at the position that precedes a low-abundance nucleotide; the entire {DNA} can therefore be sequenced in four experiments, in each of which a different type of nucleotide {(A,} G, T or C) is used in limiting concentration. Whereas issues relating to transcriptional pauses can be addressed using multiple experiments, the parallelization of this technique remains a challenge. Nanofabricated pores that are contained in biological or inorganic membranes might allow for the sequencing of individual molecules of {DNA.} As the molecule translocates across the membrane [24], the identity of individual nucleotides in the molecule of {DNA} is determined using ionic current (the flow of positive ions in the opposite direction to the negatively charged {DNA} translocating the nanopore) or transverse current (the current that flows perpendicular to the direction that {DNA} translocates) [25] {(Figure} 2). The major challenges facing this method include (i) translocating individual molecules of {DNA} across the membrane, (ii) maintaining the {DNA} in an unfolded state during translocation, and (iii) defining a general strategy for the identification of individual bases during translocation. These issues have been addressed in a number of studies. Dekker et al. achieved the translocation of a 97 kbp-long {DNA} molecule across a nanopore using electrophoresis [26]. Li et al. reported a majority of unfolded molecules of {DNA} translocating ion-beam fabricated nanopores [27]. Martin et al. fabricated conical nanopores that can detect analytes with greater sensitivity than cylindrical nanopores [28]. The use of engineered hairpins to ‘stall’ the translocation of {DNA} revealed single-nucleotide differences between hairpins of the same length [29]. An important issue in deciphering bases electronically is the identification and removal of the sources of electronic noise. Much work remains to be done to identify unequivocally the sources of measured noise in a nanopore experiment. A potential challenge for reducing this noise is the need for the experiments to be done at room temperature. Indeed, reduction of electronic white noise is commonly achieved by performing measurement at cryogenic temperatures. Unfortunately, the need to translocate {DNA} through a nanopore in a solution is incompatible with this noise-reduction method. Coding of the read-out signal (similar to the methods used in telecommunication channels for reliable data transmission) is another method, as yet unexplored, to address the reduction of electronic noise. Recently, an encouraging study has shown that the electronic identification of individual ribonucleosides using hemolysin pores that contain a β-cyclodextrin molecule {(Figure} 3) [30]. This method might be able to identify bases that are cleaved from {DNA} using an exonuclease; however, the rate of cleavage must be slower than the rate of identification and a complete sequencing scheme must be structured around this detection method. The ability to observe the addition of a single nucleotide to a molecule of {DNA} would provide a radically new way to sequence {DNA.} The observation of single molecular events typically requires nano- to picomolar concentrations to isolate individual molecules; these concentrations, however, are below the micromolar concentration range at which most biological processes occur. Webb et al. have generated an array of subwavelength apertures (ranging from 50 nm to 200 nm in width) in a thin metal film and, using the evanescent light emitted from the metal apertures, observed the addition of dye-labeled {dCTP} to {DNA} at micromolar concentrations [31]. The parallel nature of this experiment indicates the potential for observing simultaneously many single-nucleotide additions, and might provide the basis for a new sequencing technology. One challenge faced by this technology is the need to isolate single molecules in a well because the presence of two or three molecules of {DNA} will produce non-uniform signals. The performance of established and emerging {DNA} sequencing methods are quantified in Table 1 according to cost, speed, accuracy and read length. In addition, the simulated performance of a {‘US\$1000-genome} sequencing technology’ is also estimated. The concept of a {US\$1000-genome} sequencing technology is an idealized and ambitious objective, which aims to develop instruments that allow for the sequencing of {DNA} relying on cheaper and faster methods than those currently in use, while maintaining their accuracy. In this section, we discuss the potential of these technologies to meet the criteria for such technology. A {US\$1000-genome} sequencing technology requires a 104-fold reduction in cost per base and a 103-fold increase in the number of bases read per second. The accuracy and read length obtained using established and emerging methods approximate those required of future genome sequencing technologies, however, it is unclear whether parallelization of these methods alone will provide the necessary improvements in cost and throughput required for a {US\$1000-genome} sequencing technology. The technologies that sequence single molecules of {DNA} directly are unknown entities with regard to cost and we can therefore only speculate that their development might yield a {US\$1000-genome} sequencing technology. We anticipate that the use of cheaper materials (e.g. plastics in place of glass for miniaturized sequencing devices [12] and [32]), the availability of rapid prototyping for manufacturing miniaturized devices, and the reduction in reagent volumes might reduce the cost to {US\$100} 000 per genome. This is a ‘near-term’ cost target set by the National Institutes of Health {(NIH)} for the year 2010, and we believe that this target can be met using existing technology, albeit with considerable development and clever use of materials. The over-arching goal of sequencing whole genomes for {US\$1000} hence remains a challenge for basic and applied science. Conventional Sanger sequencing remains the method of choice for genome analysis. However, the miniaturization of this method using microfluidics and microelectrophoresis can present significant cost reductions (on a per read basis). It is clear that the serial loading of microchannels for sample purification will limit the application of miniaturized Sanger sequencing technologies, and the Massively Parallel Strategies that exploit clonal amplification might prevail in future applications. Single-molecule sequencing methods that rely on the measurement of the motion of a polymerase, the tunneling current across {DNA} or the fluorescence from an individual nucleotide represent radically new approaches to identify nucleotides. The methods are, however, at an early stage and the problems associated with, for example, rendering single-molecule displacement measurements parallel or guaranteeing that an individual molecule (as opposed to two or more molecules) resides in a single well pose significant challenges. Our survey shows that the development of unconventional and exciting approaches to sequence {DNA} using nanofabrication might provide the revolutionary advances that would allow the sequencing of whole genomes cheaply and quickly; it is probable that the ability to fabricate smaller structures that exploit electronic (e.g. tunneling current) and optical (e.g. zero-mode waveguides) phenomena will be central to new sequencing technologies. It remains unclear, however, whether the methods currently available will provide the basis for a {US\$1000-genome} sequencing technology. The authors gratefully acknowledge financial support from the Gordon and Betty Moore Foundation and the National Academies Keck Future Initiative. },
	number = {9},
	journal = {Trends in Biotechnology},
	author = {Declan Ryan and Maryam Rahimi and John Lund and Ranjana Mehta and Babak A. Parviz},
	month = sep,
	year = {2007},
	pages = {385--389}
},

@article{welch_syntheses_1999,
	title = {Syntheses of Nucleosides Designed for Combinatorial {DNA} Sequencing},
	volume = {5},
	url = {http://dx.doi.org/10.1002/(SICI)1521-3765(19990301)5:3<951::AID-CHEM951>3.0.CO;2-G},
	doi = {{10.1002/(SICI)1521-3765(19990301)5:3{\textless}951::AID-CHEM951{\textgreater}3.0.CO;2-G}},
	abstract = {Fluorescent and photolabile nucleotide analogues 1 and 2 were prepared to test an unconventional approach to {DNA} sequencing that, ideally, would not involve gel electrophoresis. These particular compounds were not incorporated, but modeling studies indicate how further refinements to their structures could facilitate the proposed sequencing scheme. B = adenine, thymine; {TP} = triphosphate,},
	number = {3},
	journal = {Chemistry - A European Journal},
	author = {Mike B. Welch and Carlos I. Martinez and Alex J. Zhang and Song Jin and Richard Gibbs and Kevin Burgess},
	year = {1999},
	pages = {951--960}
},

@article{shapir_puzzle_2005,
	title = {The puzzle of contrast inversion in {DNA} {STM} imaging},
	volume = {109},
	number = {30},
	journal = {Journal of Physical Chemistry {B-Condensed} Phase},
	author = {E. Shapir and J. Yi and H. Cohen and A. B. Kotlyar and G. Cuniberti and D. Porath},
	year = {2005},
	pages = {14270--14274}
},

@article{postma_rapid_2008,
	title = {Rapid Sequencing of Individual {DNA} Molecules in Graphene Nanogaps},
	journal = {Arxiv preprint {arXiv:0810.3035}},
	author = {H. W. C. Postma},
	year = {2008}
},

@inproceedings{hansma_progress_1991,
	address = {Baltimore, Massachusetts {(USA)}},
	title = {Progress in sequencing deoxyribonucleic acid with an atomic force microscope},
	volume = {9},
	url = {http://link.aip.org/link/?JVB/9/1282/1},
	doi = {10.1116/1.585221},
	booktitle = {Proceedings of the Fifth International Conference on Scanning Tunneling {Microscopy/Spectroscopy}},
	publisher = {{AVS}},
	author = {H. G. Hansma and A. L. Weisenhorn and S. A. C. Gould and R. L. Sinsheimer and H. E. Gaub and G. D. Stucky and C. M. Zaremba and P. K. Hansma},
	month = mar,
	year = {1991},
	keywords = {{ATOMIC} {FORCE} {MICROSCOPY,BIOPHYSICS,DNA} {SEQUENCING,NUCLEOTIDES,RESOLUTION,SAMPLE} {PREPARATION,USES}},
	pages = {1282--1284}
},

@article{hansma_atomic_2004-1,
	title = {Atomic force microscopy imaging and pulling of nucleic acids},
	volume = {14},
	issn = {{0959-440X}},
	url = {http://www.sciencedirect.com/science/article/B6VS6-4CDS7DK-4/2/b48944ee3d8a5e8a4d731dc885be8f73},
	doi = {10.1016/j.sbi.2004.05.005},
	abstract = {
Recent advances in atomic force microscopy {(AFM)} imaging of nucleic acids include the visualization of {DNA} and {RNA} incorporated into devices and patterns, and into structures based on their sequences or sequence recognition. {AFM} imaging of nuclear structures has contributed to advances in telomere research and to our understanding of nucleosome formation. Highlights of force spectroscopy or pulling of nucleic acids include the use of {DNA} as a programmable force sensor, and the analysis of {RNA} flexibility and drug binding to {DNA.}},
	number = {3},
	journal = {Current Opinion in Structural Biology},
	author = {Helen G Hansma and Kenichi Kasuya and Emin Oroudjev},
	month = jun,
	year = {2004},
	pages = {380--385}
},

@article{lindsay_canscanning_1991,
	title = {Can the scanning tunneling microscope sequence {DNA?}},
	volume = {8},
	issn = {1050-3862},
	url = {http://www.ncbi.nlm.nih.gov/pubmed/2043383},
	abstract = {A revolutionary new microscope, the scanning tunneling microscope {(STM),} can image some surfaces at atomic resolution, even in air or water. It can produce high-resolution images of {DNA,} and we outline what we know of its mechanism, concluding that it may be able to sequence {DNA.} This application would require major advances in sample preparation in order for the technique to compete with conventional methods. On the other hand, the {STM} may provide a very useful alternative to gels for probing sequence-directed structural features.},
	number = {1},
	journal = {Genetic Analysis, Techniques and Applications},
	author = {S M Lindsay and M Philipp},
	month = feb,
	year = {1991},
	note = {{PMID:} 2043383},
	keywords = {Base {Sequence,Microscopy,} Scanning {Tunneling,Nucleotides,Polydeoxyribonucleotides}},
	pages = {8--13}
},

@article{driscoll_atomic-scale_1990,
	title = {Atomic-scale imaging of {DNA} using scanning tunnelling microscopy},
	volume = {346},
	url = {http://dx.doi.org/10.1038/346294a0},
	doi = {10.1038/346294a0},
	number = {6281},
	journal = {Nature},
	author = {Robert J. Driscoll and Michael G. Youngquist and John D. Baldeschwieler},
	month = jul,
	year = {1990},
	pages = {294--296}
},

@article{chang_tunnelling_2009,
	title = {Tunnelling readout of hydrogen-bonding-based recognition},
	volume = {advanced online publication},
	issn = {1748-3395},
	url = {http://dx.doi.org/10.1038/nnano.2009.48},
	doi = {10.1038/nnano.2009.48},
	journal = {Nat Nano},
	author = {Shuai Chang and Jin He and Ashley Kibel and Myeong Lee and Otto Sankey and Peiming Zhang and Stuart Lindsay},
	month = mar,
	year = {2009}
},

@article{xu_electronic_2007,
	title = {The Electronic Properties of {DNA} Bases},
	volume = {3},
	url = {http://dx.doi.org/10.1002/smll.200600732},
	doi = {10.1002/smll.200600732},
	abstract = {No Abstract},
	number = {9},
	journal = {Small},
	author = {Mingsheng Xu and {Robert�G.} Endres and Yasuhiko Arakawa},
	year = {2007},
	pages = {1539--1543}
}