Sequencing, Assembling, and Annotating a Mid-Sized Genome

Quality of genome assembly is a crucial factor affecting gene discovery. Hybrid genome assembly approaches aim to simultaneously use data from multiple high-throughput sequencing platforms to produce improved assembly. Second generation sequencing platforms can produce short accurate reads at a high depth of coverage, but these are restricted by inability to span long repeats. On the other hand, third generation platforms can produce long reads with high error rate at higher costs whereas jumplng libraries have high variability in insert size.

Hybrid assembly approaches have produced high quality assemblies for simple bacterial genomes which have low repeat percentage and can be trivially sequenced at high coverage. However, current automated assembly tools are still testing the waters in producing sufficiently high quality assembly for repeat-rich mid-sized genomes to enable gene discovery. The user may not always have the freedom to select the right mix of sequencing data and the assembly software should have the flexibility to generate the best assembly with available data.

In this work, we assemble mid-sized genomes using our Cerulean hybrid genome assembly approach. This approach starts with an assembly graph generated using the short second generation reads and iteratively refines the assembly graph using data sets with longer reads or insert sizes. We characterize how the assembly quality changes with availability of data, especially with the aim to facilitate the practical problem of gene discovery.