Recent advances in DNA sequencing methods – …

1 ReviewRecent advances in DNA sequencing methods general principlesof sample preparationSten LinnarssonLaboratory for Molecular Neurobiology, Department of Medical Biochemistry and Biophysics, Karolinska Institutet, Scheeles v g 1,SE-171 77 Stockholm, SwedenARTICLE INFORMATIONABSTRACTA rticle Chronology:Received 12 February 2010 Accepted 28 February 2010 Available online 6 March 2010 DNA sequencing has revolutionized biomedicine, and progress in the field has been unrelentingsince it was invented over 30 years ago. The complete DNA sequence of the human genome wasobtained as the culmination of a decade of work by a large number of scientists.

2 Less than ten yearslater, so-called next-generation instruments now make it possible for a single lab to produce thesame amount of data in a week. But while the instruments are increasingly automated, upstreamsample processing remains a challenge. Here I review the current state of the art in preparinggenomic and RNA samples for high throughput sequencing . 2010 Elsevier Inc. All rights :DNA sequencingSample for a sequencing -ready sample preparation end-game: quality control and RNA for A. Supplementary since DNA sequencing was invented in the mid-seventies byFredrick Sanger (the chain-terminating method[1]) and by AllanMaxam and Walter Gilbert (the chemical method[2]), sequencinghas been a solid foundation for research in all branches of biologyand medicine.

3 In Recent years, so-called next-generation sequen-cing instruments have been developed, which are in general basedEXPERIMENTAL CELL RESEARCH 316 (2010) 1339 1343E-mail see front matter 2010 Elsevier Inc. All rights at a clonal amplification method to directly generate clustersof DNA templates at high densities on a glass slide. Templatesare then sequenced by a stepwise incorporation of nucleotides(Illumina Genome Analyzer, 454 Genome Sequencer) or shortoligonucleotides (Applied Biosystems SOLiD). Through impressiveincremental improvements the throghput of these instrumentshas increased to currently about 200 gigabases per week perinstrument (Illumina HiSeq 2000), while costs have dropped towhere a human genome can be sequenced at 30-fold coverage forless than $10,000.

4 These developments have recently been thesubject of several excellent reviews[3 5].However, while the actual sequencing has been streamlinedand automated, the upstream sample preparation procedureremains an area of active and fertile research. For every applicationof DNA sequencing , such as metagenomics, transcriptome se-quencing (RNA-Seq), genome sequencing or resequencing, chro-matin immunoprecipitation sequencing (ChIP-Seq) and so on,there is the need for a specific protocol to convert the sourcematerial into a standard DNA library suitable for loading onto thesequencing I will review Recent developments in sample preparationwith a view to providing some useful advice to the researcher whowishes to convert any kind of raw nucleic acid into a sequencing -ready library.

5 The focus will be on sample preparation for thecommercial platforms already on the market, in particular forthe 454 Genome Sequencer FLX (henceforth termed the FLX ), theIllumina Genome Analyzer IIx ( GA , used loosely to include therecent HiSeq 2000) and the Applied Biosystems SOLiD 3 Plus( SOLiD ), although many general principles will apply also toemerging and future systems. Enrichment methods , used to selec-tively enrich desired target regions in a genome, have beenreviewed elsewhere[3,6]and will not be covered here. Finally,I will briefly review the way these general principles apply toone specific objective, sequencing RNA.

6 In the accompanyingSupplementary methods , a detailed best-practices protocol isprovided, as currently implemented in our for a sequencing -ready libraryFor each of the current platforms, the input is a double-strandedDNA library consisting of short fragments flanked by adapters ofknown (and platform-specific) sequence. Given the great differ-ences between the GA, FLX and SOLiD platforms in both how theygenerate clonally amplified templates on the flowcell surface, andhow these clones are interrogated to reveal their sequence, theinput sample requirements are remarkably similar (Table 1). Theprincipal difference is that FLX and SOLiD clonally amplify thetemplates by emulsion PCR[7], whereas the GA uses surface PCR,also known as bridge amplification[8].

7 The SOLiD system uses anemulsion with smaller droplets compared with FLX, and worksbest with shorter fragments. In contrast, the GA accepts a greaterrange of fragment lengths, but the yield drops as fragments getlonger because longer fragments result in larger surface clustersthat must be spaced less 's instructive to consider how much sequence information iscontained in a typical sequencing library. As an example, 2 L of 2nM double-stranded DNA with an insert length of 300 bp billion molecules weighing less than one nanogram. The 720gigabases of sequence contained in these molecules are equivalentto about one hundred human diploid genomes.

8 This is morethan enough complexity for current instruments, but not by agreat margin; the most Recent GA instruments can generate 200gigabases per we can see that the minimum unamplified startingmaterial must be about one nanogram. For genome sequencing , atleast a hundred diploid cells would be required if there were nolosses in sample preparation. However, if parts of the genome aretargeted using enrichment methods , then proportionately largeramounts of starting material will be required. Similarly, sincemRNA comprises only about 300,000 molecules per mammaliancell, each about 2 kb long, more than a thousand cells would berequired even if mRNA could be converted to a sequence-readylibrary without present, therefore, all sample preparation protocols for next-generation sequencing require either a large amount of startingmaterial (which blurs distinctions between individual cells), orsome form of amplification (which introduces inevitable bias), orboth.

9 Recent developments in sample preparation have sought toreduce the impact of both of these sample preparation proceduresIn the general case, common to almost all sample preparationprotocols for next-generation sequencing , the starting material isdouble-stranded DNA in the form of isolated genomic DNA,reverse-transcribed cDNA, or immunoprecipitated chromatin. Toconvert this into a sequenceable library, the source DNA must befragmented, polished, size selected, adapter ligated, purified most cases there will be no need for specialized , great care should be taken to avoid cross-contamination,as this will be very difficult to detect.

10 In addition, wheneverworking with low concentration of nucleic acid, it is necessary touse low-adsorbing plasticware. We and others[9]have found thatstandard polypropylene microfuge tubes can adsorb as much as100 ng DNA, which can lead to complete loss of a sample. Werecommend using polyallomer tubes (Beckman Coulter; non-stick tubes from other manufacturers may be equivalent) and toadd some detergent in every reaction step ( Tween-20)to reduce adsorption to tube is typically achieved by mechanical force. Theofficial protocols provided by the manufacturers call for fragmen-tation by nebulization using a disposable nabulizer driven bypressurized air (GA, FLX), or by the Covaris AFA ultrasound device(SOLiD).