Main

454 Life Sciences has developed and commercially released a new protocol for generating a library of paired-end fragments to determine the orientation and relative positions of contigs produced by de novo shotgun sequencing and assembly. This 3K Long-Tag Paired End protocol (Fig. 1) can also be used to identify genomic structural variations1 and their associated breakpoints. Structural variation of the genome, involving large, kilobase- to megabase-sized deletions, duplications, insertions, inversions and complex combinations of rearrangements, is widespread in humans and is presumably responsible for a considerable amount of phenotypic variation. The 3K Long-Tag Paired End library DNA fragments comprise an approximately 250-bp fragment with a 44-mer adaptor sequence in the middle, flanked by 100-mer sequences, on average. The two flanking 100-bp sequences are segments of DNA that were originally located approximately 3 kb apart in the genome of interest.

Figure 1: A schematic of the 3K Long-Tag Paired End sequencing protocol.
figure 1

SA, streptaviden; Met, methylated; Bio, biotin.

Full size image

Traditional approaches to the sequencing of paired-end reads rely upon inserting a DNA fragment into a vector, such as a bacterial artificial chromosome or a fosmid, cloning this into bacteria and subsequently generating two sequences, one from each end of the vector. These methods entailed weeks of laboratory work and could cost several hundred thousand dollars to prepare the libraries needed for Sanger sequencing. The Genome Sequencer FLX method presented here, which requires no cloning, generates up to 200,000 paired-end reads from a single Genome Sequencer FLX instrument run with a total elapsed time—from genomic DNA to result—of less than 4 days.

Sample preparation protocol

The preparation of a 3K Long-Tag Paired End library is depicted schematically in Figure 1. The protocol begins with fragmentation of the high-molecular-weight DNA sample; the size distribution of the fragments (on average 3 kb) determines the distance between the paired-end sequencing tags. The fragments are methylated to prevent EcoRI cleavage, Hairpin Adaptors (biotinylated and containing non-methylated EcoRI recognition sites, provided in the GS Paired End Adaptor Kit) are ligated onto both ends, and all DNA species that are not protected by hairpins are removed by exonuclease digestion. The remaining long insert fragments are circularized by digestion with EcoRI to remove the terminal hairpin structures, providing cohesive ends for ligation. The resulting 3-kb circular fragments contain the 44-bp linker (the remainder of the two Hairpin Adaptors) joining the two ends of the fragmented DNA.

The DNA circles are then fractionated by nebulization, generating molecules that are a few hundred base pairs in length. Long Paired End Adaptors are ligated to the ends of the linker-positive fragments. The adaptors provide priming sequences for both amplification and sequencing of the Paired End library fragments. This library is ready for emulsion-based clonal amplification (emPCR™) using the GS emPCR Kit II (Amplicon A, Paired End) and for sequencing using appropriate GS Sequencing and GS PicoTiterPlate kits and the Genome Sequencer FLX instrument.

Assemblies

Examples of assemblies resulting from the 3K Long-Tag Paired End protocol are shown in Table 1. We assembled three bacterial genomes, Escherichia coli K-12, Thermus thermophilus and Campylobacter jejuni, by three different sequencing methods: only shotgun sequencing reads (250–300 bases in length); a 50:50 mix of shotgun and 3K Long-Tag Paired End sequencing reads; and only 3K Long-Tag Paired End sequencing reads. In all of the assemblies, the number of reads in the data sets was minimized to an approximately 20× depth of coverage by randomly discarding sequence reads. The data used in the 50:50 mix were a 10× depth of shotgun reads and 10× depth of reads generated using the 3K Long-Tag Paired End protocol.

Table 1 Comparison of data from three methods of de novo assembly
Full size table

The GS de novo Assembler Software (version 1.1.03) identifies the reads as either linker positive or linker negative. The initial step in the assembly is the generation of a de novo shotgun assembly using the linker-negative reads and the DNA reads on either side of the linker. Once the de novo assembler places the shotgun reads into contigs, the linker-positive reads (Long-Tag Paired End reads) are used to orient the contigs into scaffolds (Fig. 2). This assembly method is used with 3K Long-Tag Paired End read data alone and when the 3K Long-Tag Paired End data are mixed with shotgun data.

Figure 2: Sequencing results from a typical 3K Long-Tag Paired End library prep of E.coli K-12.
figure 2

A region of the de novo assembly of E. coli K-12, with the de novo–assembled contigs covering the region shown in blue along the bottom axis. The paired-end reads generated with this protocol are capable of bridging the 0.2-kb and 1.5-kb gaps between the contigs, highlighted in green.

Full size image

All three methods of assembly generate comprehensive, highly accurate assemblies (Table 1). The choice of which experimental approach and assembly method to use depends on the goals of the research. If a quick view of the genome (for example, to identify which genes are present) is desired, a shotgun-only approach is suitable. If the research goal is to generate a high-quality draft of the target genome, then the inclusion of Long-Tag Paired End data is the best option.

Summary

The sequencing of kilobase-sized inserts is quite valuable for a number of applications, including improved de novo assembly and identification of genomic structural variations. The 3K Long-Tag Paired End protocol provides a quick, efficient and cost-effective method for generating hundreds of thousands of sequence reads, each containing a pair of ∼100-bp reads separated by 3-kb size inserts. Future development plans include a protocol for sequencing tags separated by 15- to 20-kb distances. The combination of both 3-kb and longer paired-end spacing will better enable the assembly of larger and more complex genomes.

Additional information about the Genome Sequencer System is available from Roche Applied Science (http://www.genome-sequencing.com). 454, 454 Life Sciences, 454 Sequencing, emPCR and PicoTiterPlate are trademarks of 454 Life Sciences Corporation, Branford, Connecticut, USA. For life science research use only. Not for use in diagnostic procedures. License disclaimer information is available online (http://www.genome-sequencing.com).