The US Department of Energy's Joint Genome Institute is a high-throughput sequencing center and user facility that has sequenced a large number of microbial genomes. The strategy for most projects calls for construction of whole genome shotgun libraries from high-molecular weight DNA isolated from an axenic culture. In general, the JGI produces 3 insert size-selected libraries for all whole genome shotgun projects. We generate a 3kb high-copy pUC18 library, an 8kb low-copy pMCL200 library, and a 40kb pCC1FOS fosmid library. The DNA is randomly sheared, fragments are end-repaired for blunt-end cloning, and then size selected on an agarose gel, extracted and purified. 3 & 8kb inserts are cloned into the appropriate vector and transformed into E. coli. 40kb inserts are cloned, packaged and infected by phage into E. coli. PCR using primers flanking the inserts are used to determine the percentage of clones with inserts for both the 3 and 8kb libraries, before proceeding to production sequencing. Clones (10-384-well plates) from each of the 3 & 8kb libraries are initially sequenced and library quality is assessed at this stage before full sequencing is completed. Both 3 & 8kb libraries are sequenced to 4x sequencing coverage and the 40kb library is sequenced to 30x clone coverage. The 3 library approach generally results in more complete genome coverage at the draft stage, and pairing information allows for contig order and orientation and repeat resolution in the sequence. Finishing using standard methods is also facilitated by this approach.

As Sanger sequencing is being replaced by higher throughput and lower cost of next generation sequencing, finishing microbial genomes will face two major challenges. First, the technology will need to be fast enough to handle many more drafted genomes. Second, it will have to incorporate a clone-free approach to fill gaps. We have been testing a method that utilizes a universal "bubble-tag" to perform primer walking and gap closure in a clone-free condition. The ?bubble-tag? method was first described by Doug Smith (PCR Methods Appl. 2: 21-27, 1992) to amplify and sequence lambda DNA. There is no evidence however, that this approach will work for the more complex microbial genome. Here we describe the experimentation of this approach in primer walking of the E. coli genome. Genomic DNA was sheared, blunt-end repaired, or digested with frequent cutters, and ligated to bubble adaptors. Site specific primers were used together with the universal bubble primer to amplify and sequence the regions of interest. Different primer concentration was tested in order to generate long amplified templates. This approach enables primer walking and gap filling in a clone-free draft sequencing process. More importantly, the uniformity of this approach is amenable for an automated finishing process.