Analysis of Illumina's Nextera Long Mate Pair Data, and Its Effects on Sequence Assembly

The whole genome shotgun assembly problem, especially when using data from next generation sequencing, is impeded by repeats longer than the sequenced fragments. While repetitions larger than the effective read length can be resolved by using paired end sequencing, the creation of long paired libraries in the range of thousands of base pairs in fragment size is more challenging than simple paired end libraries, introduces more bias, and requires the use of more complex analyses and algorithms.

The current version of Long Mate Pair library construction protocol from Illumina, Nextera, generates reads with a linker sequence that can be used to detect true-mates from contamination. We have previously shown how the classification and cleanup of true-mates enhances the scaffolding results. We are now analysing how these cleaner, better libraries can be used both in the scaffolding step and previous contig construction steps, by using more strict analysis, filtering and classification steps based on the fragment properties and linker sequence to prepare the reads.

We present here the current version of our methods to use Nextera libraries, both combined with standard and PCR-free paired end data, and on its own, to perform de novo assembly and scaffolding. We also show the key QC steps we take into account when evaluating these libraries, and analyse their final performance when used with different assemblers and scaffolders, compared to the expected effect predicted by mapping to reference genomes