Genome Assembly of an Amazonian Elaeis oleifera Genotype and Comparative Analysis with Other Oil Palm Genomes

Although Elaeis oleifera (Eo) is a source of variability for traits of interest in Oil palm (E. guineensis - Eg) breeding programs, it lacks the genomic resources currently available for Oil Palm. Therefore, we sequenced the genome of a native Eo genotype in nine lanes of Illumina Hiseq2000, generating 2.1 billion reads. After quality control (phred 20, 80%), 2 billion reads remained to be further analyzed. Considering an assembly estimate size of 1,5 Gbp, we ended up after QC with: ~115X of 180bp (insert length) and ~14X of 800bp paired-end libraries; ~7X of 3kbp, ~4X of 5kbp and ~3X of 10kbp mate-pairs libraries. The assembly of the ~137X cleaned data using the AllPaths-LG assembler resulted in ~85k scaffolds, with a N50 of 27kbp. The addition of an extra 0.8X generated using 454 long inserts (8kbp and 20kbp) through a hybrid assembly strategy using MIRA and SOAPdenovo did not have any significant impact on assembly quality. We also compared our draft sequence with public available Eg and Eo genomes using MUMer. We found that our draft covers ~70% of the first genome, and ~78% of the latter. Considering the cost difference between Illumina HiSeq and Roche 454, we are confident that we managed to obtain good quality genome sequence spending much less. Our work now is concentrated in three directions: (i) gene prediction for all three genomes, (ii) increase of sequence depth using Illumina large inserts jumping libraries to improve scaffolding and (iii) genetic mapping.