An Optimised Approach to Sequence and Assemble BACs: Application to Bread Wheat Chromosome Arm 7DS

We are currently sequencing a minimal tiling path of BACs for wheat chromosome arm 7DS. This project uses an in house developed assembler ‘Sassy’, which has been designed specifically for complex BAC assembly and requires more data and memory than traditional assemblers, but produces better assemblies. Sassy performance was demonstrated in a recent publication by the assembly of organelle genomes.

Optimisation of the data generation is a critical aspect of producing the best assemblies with Sassy. We currently sequence four (preferably non-overlapping) BACs in a pool, and run 96 pools (384 BACs) indexed on one lane of HiSeq. This generally provides > 500x coverage for each BAC with 100 bp paired reads. This pooling/indexing strategy was optimised to generate the amount of data Sassy requires for assembly. DNA was normalised to limit bias. The DNA library preparation was also optimised and truSeq kits were shown to outperform the cheaper and quicker Nextera methods.

Following sequencing, the resulting data is deconvoluted back to the 96 pools of four BACs. QC for data quality and quantity is performed and the data is then assembled using Sassy. Assemblies are still being optimised but current results suggest that BACs are represented by between one and seven contigs, with an average of 2.7 contigs per BAC.

For each pool, BAC end sequences are mapped to the contigs to identify the ends of BAC assemblies and to link contigs to specific BACs. Whole genome Illumina mate pair libraries have been generated using the new Illumina MP kit. The resulting data are mapped to each BAC pool and used to scaffold the contigs to complete BACs. This, together with 7DS PACBio data, has been used to confirm the quality of the assemblies.