P1176
Assembling and Comparing Two Ash Tree Genomes: Coping with Heterozygosity on Non-Model Organism WGS

Bernardo Clavijo , The Genome Analysis Centre, Norwich, United Kingdom of Great Britain and Northern Ireland
Gonzalo Garcia , The Genome Analysis Centre, Norwich, United Kingdom of Great Britain and Northern Ireland
Jonathan M Wright , The Genome Analysis Centre, Norwich, United Kingdom of Great Britain and Northern Ireland
Darren Heavens , The Genome Analysis Centre, Norwich, United Kingdom
Matt Clark , The Genome Analysis Centre, Norwich, Norfolk, United Kingdom
Federica Di-Palma , The Genome Analysis Centre, Norwich, United Kingdom of Great Britain and Northern Ireland
Sequencing and assembling biologically interesting samples often means leaving the easy single-reference, model organism, genome scenario towards a myriad of interesting but extremely challenging individual genomes. On this work we focus on two different Ash tree (Fraxinus excelsior) genomes, chosen from a sample group exposed to the Ash Dieback disease. This two trees, one of them less susceptible to the disease, where highly heterozygous, and DNA extraction proved challenging.

We created a genome sequencing and assembly strategy based on Illumina PCR-free Paired End libraries and Nextera Long Mate Paired libraries, two techniques that would produce good data and would be achievable with our starting material. Typical pipelines produce assemblies with poor contiguity, duplicated sequence and most worryingly even loss of homozygous content. We designed a post-contigging filter, using kmer-spectra properties to keep all the homozygous content and discard roughly half of the heterozygous content in the final set of contigs. This collapsing simplifies the scaffolding problem downstream reducing the number of alternative paths for the use of long range information, which in turn produces a higher quality and more contiguous assembly with all the content for a complete collapsed mosaic.

This sequencing and assembly approach created two comparable assemblies, working robustly and consistently on both samples, allowing us to carry on towards downstream analysis. We present here both the method for heterozygous assemblies and the results of some preliminary comparative analyses based on the genome sequences, to showcase the utility of assemblies created with this technique.