Assemblly Of The 24Gb Loblolly Pine Genome From WGS Short Read Data

We present the Super Reads approach that we used to assemble 24Gb (estimated) Loblolly Pine genome from short-read Illumina data.  The main challenge of this project was to assemble a whole genome shotgun data set of over 15 billion short 100-160bp reads.   This task has never been undertaken before and seemed impossible only a few years ago.

The super reads approach has the computational efficiency of the de Bruijn graph and the robustness of the overlap-based assembly strategies. It is uniquely suited to handle the rich repeat families found in plant genomes. In this approach we transform the huge number of paired end Illumina reads into a much smaller number of longer Super-reads. For the Loblolly Pine data set we reduced the 14B short paired end reads to 146 Million super reads containing 52Gb of sequence.  Half of that sequence was in super reads of 500bp or longer.

The combined set of the super reads and the filtered set of jumping library and DiTag mate pairs providing connectivity information contained about 636M reads.  We then used a modified version of CABOG assembler based on the 6.1 release to create the final genome assembly. We target at least 5-10kb N50 contig size and over 1Mb N50 scaffold size for the final assembly.