Assembly Strategy for the BAC Pool Sequencing of Aegilops tauschii, the Ancestor of the Wheat D Genome

Previous efforts to sequence wheat genomes using the whole-genome shotgun (WGS) method have resulted in fragmented assemblies, much smaller than the genome size. This is likely due to the unusually high repetitive content of the wheat genome. The bread wheat species, Triticum aestivum, is an allohexaploid, making assembly especially difficult. T. aestivum comprises three ancestral genomes, of which one, wheat D, is a very close relative of the wild grass Aegilops tauschii. The spontaneous hybridization of Ae. tauschii and a recent wheat ancestor produced the modern hexaploid T. aestivum genome. An assembly of Ae. tauschii, therefore, provides a very high-fidelity genetic model for one of the three component genomes of bread wheat.

We are sequencing the 4.3 gigabase Ae. tauschii genome using pools of 6 to 10 overlapping BACs. Each BAC is approximately 145 Kb in length, while the pools size span ~1Mb. As of October 2014, we have completed sequencing over 4,000 pools out of a total of 5,500 (75%), generating over 100x coverage using Illumina MiSeq paired-end libraries. We have completed all pools from chromosomes 1D to 5D, and chromosomes 6D and 7D are in progress. We had previously sequenced chromosome 3DS (the short arm of chr 3) using non-overlapping BACs and 454 sequencing technology.

We used SOAPdenovo2 to generate the original BAC pool assemblies. In addition to the pair-end libraries, we used publicly available WGS long-insert libraries for scaffolding.¹ We implemented new methods for WGS mate-pair selection, long terminal repeat assembly, and assembly of low-coverage regions especially for this project. Using these methods, we have been able to increase the scaffold N50 size by more than 50%, from ~133 Kb (produced by SOAPdenovo2) to ~205 Kb. We have validated the assemblies using long-insert paired sequences, BAC end sequences, and alignments from a restriction map produced using BioNano Genomics mapping technology.

We will present these methods along with preliminary results on merging the pools using the minimus² assembler. This work is a part of NSF-funded Project IOS-1238231 to generate a reference sequence for the genome of Ae. tauschii (http://aegilops.wheat.ucdavis.edu/ATGSP/).

1. J. Jia et al. (2013). Aegilops tauschii draft genome sequence reveals a gene repertoire for wheat adaptation. Nature 496, 91-95.

2. T. Treangen et al. (2011). Next geneation sequence assembly with AMOS. Curr Protoc Bioinformatics. 2011 Mar;Chapter 11:Unit 11.8. doi: 10.1002/0471250953.bi1108s33.