A scalable, fully automated process for construction of sequence-ready barcoded libraries for 454

The emergence of next-generation sequencing technologies, such as the Roche/454 Genome Sequencer, the Illumina Genome Analyzer, the Applied Biosystems SOLiD sequencer and others, has provided the opportunity for both large genome centers and individual labs to generate DNA sequence data at an unprecedented scale [1]. However, as sequence output continues to increase dramatically, processes to generate sequence-ready libraries lag behind in scale. The minimum unit of sequence data (for example, lane or channel) already exceeds the amount required for small projects, such as viral or bacterial genomes, and will continue to increase. As a result, projects with large numbers of samples but small sequence per sample requirements become increasingly challenging to undertake in a cost-effective manner.The 454 Genome Sequencer uses bead-in-emulsion amplification and a pyrosequencing chemistry to generate DNA sequence reads by synthesis [2]. Longer reads and shorter sequencing run times make the 454 platform a powerful tool for de novo assembly of small genomes, metagenomic profiling and amplicon sequencing compared with other next-generation sequencing platforms. However, these types of applications pose a challenge in that they require a relatively small number of reads from large numbers of samples. For example, for viruses such as HIV, the small (approximately 10 kb) genome size means that a single sample on even the smallest scale 454 picotiter plate configuration (1 region of a 16 region gasket) would yield over 1,500-fold coverage, vastly more coverage than required for genome assembly. Further, the standard 454 library construction protocol is not easily scalable and becomes a major cost driver relative to sequencing when modest numbers of reads are required from each sample. In addition, when sequencing large numbers of isolates of the same organism, the sequence identity between samples makes cross-contamination virtually impossible to detect without a molecular (s