Repetitive organisms pose a challenge for short read assembly, and typically only unique regions and repeat regions shorter than the read length, can be accurately assembled. Recently, we have been investigating the use of Pacific Biosciences reads for de novo fungal assembly. We will present an assessment of the quality and degree of repeat reconstruction possible in a fungal genome using long read technology. We will also compare differences in assembly of repeat content using short read and long read technology.

Sequencing-by-synthesis (SBS) technologies have provided new cost efficient ways of obtaining a high quality draft assembly as well as a fully finished genome. The result is also a change in the complex dynamics of shotgun assembly which makes it necessary to re-evaluate the whole process, especially when complete finishing is the goal. As of yet, no single SBS technology has proven itself fully capable of providing all the data necessary for efficient and streamlined finishing, therefore, a combination of SBS and Sanger data is still prefered. JGI currently utilizes both 454/Roche and Solexa/Illumina pyrosequencing in addition to the Sanger big-dye shotgun method. In this study we evaluated how various key assembly features are affected by changes in the levels of sequence from different platforms and attempted to arrive to a combination optimal for cost-efficient finishing. We also looked at how stable and predictable this optimum is in genomes with different GC and repeat composition. The criteria we picked are believed to be the best indicators, available upon the initial draft assembly, of what kind of additional effort, and how much of it, is needed to finish a microbial genome to JGI?s quality standards.

With the onset of new technology JGI has shifted its scope of work for microbes to 454 standard titanium, 454 paired end titanium, and illumina data for gap closer and quality improvement (polishing). Raw reads are assembled by Newbler to create a draft assembly that will be passed to finishers. An in-house developed software tool creates subprojects for each gap. In-silico attempts are made to close gaps using existing unassembled pyrosequence and Illumina data. Any remaining gaps are tackled by PCR based methods. These include standard PCR, bubble PCR, multiplex PCR, combinatorial PCR, and long range PCR. Once products are generated they can be sequenced, cloned or shattered as needed. Currently gap closing data is still generated using sanger. Eventually these gaps may be pooled and sequenced using illumina or 454. Once a genome is closed, illumina data is used to polish the genome. Any areas that are still substandard are subjects for resequencing. We extend this approach to fungal genomes