Anchored Assembly: A Novel Approach for Accurate and Comprehensive Variant Detection in Next-Generation Sequencing Data

Characterization of large indels, inversions, translocations, and multi-nucleotide variants is important for trait mapping and draft genome development.  These variants are largely undetected by standard pipelines. Anchored Assembly is a novel method to accurately detect large variants from next-generation sequence reads.

This method accurately detects both SNPs and structural variations with a low false positive rate and minimal reference bias.  We detect, on average, over 95% of structural variants ≤ 30 kbp, as well as 99% of SNPs. The ability to detect indels and structural variations is undiminished by the size of the variation, and the ability to accurately detect and assemble insertions continues well into the 30 kbp range.  In comparison, detection of insertions with a BWA + GATK pipeline decreases sharply above 16 bp. We assembled Illumina reads from E. coli MG1655 using an E. coli DH10B reference and detected nearly all known structural variations and SNPs, including a 56,000 bp insertion and a 144,000 bp deletion. 

Anchored Assembly is a powerful tool for accurately detecting variations on a genome-wide scale and shows potential to characterize individual and population-specific sequence regions missed by alignment to reference. This approach may be useful to assemble draft genomes using a reference of a sufficiently closely related species. We expect this method to have broad utility in the analysis of agricultural species with complex genomes, including large polyploid genomes.