Variation graph toolkit improves read mapping by representing genetic variation in the reference

Abstract

Reference genomes guide our interpretation of DNA sequence data. However, conventional linear references represent only one version of each locus, ignoring variation in the population. Poor representation of an individual's genome sequence impacts read mapping and introduces bias. Variation graphs are bidirected DNA sequence graphs that compactly represent genetic variation across a population, including large-scale structural variation such as inversions and duplications. Previous graph genome software implementations have been limited by scalability or topological constraints. Here we present vg, a toolkit of computational methods for creating, manipulating, and using these structures as references at the scale of the human genome. vg provides an efficient approach to mapping reads onto arbitrary variation graphs using generalized compressed suffix arrays, with improved accuracy over alignment to a linear reference, and effectively removing reference bias. These capabilities make using variation graphs as references for DNA sequencing practical at a gigabase scale, or at the topological complexity of de novo assemblies.

Removing reference bias and improving indel calling in ancient DNA data analysis by mapping to a sequence variation graph

“…We aligned trimmed reads to the human linear reference genome (hs37d5) using bwa aln [25] with parameters -l1024 -n 0.02 [31], keeping bases with quality above or equal to 15. We constructed the index file for vg [14] with hs37d5 and variants from the 1000 Genomes Project phase 3 dataset [13] above 0.1% MAF. In total, the graph contained 27,485,419 SNPs, 2,662,263 indels and 4,753 other small complex variants.…”

“…However, a limitation of this approach is that it has only considered biallelic single nucleotide polymorphisms (SNPs). Therefore, non-reference alleles at insertion and deletion (indel) loci are not accounted for, despite there being hundreds of thousands of non-reference indels in a typical human genome [13], and these having a greater affect on read mapping than SNPs [14].…”

“…However, the application of this approach to ancient DNA data has not yet been examined. We recently introduced the variation graph (vg) software [14], and in Figure 1 show an example of how vg can recover the alignment of short aDNA reads to alternate alleles. Here we apply vg and bwa systematically to map both simulated data and 34 previously published ancient human DNA samples, and demonstrate that mapping with vg can effectively reduce reference bias for ancient DNA samples in comparison to bwa.…”

Paragraph: A graph-based structural variant genotyper for short-read sequence data

“…Paragraph works by aligning and genotyping reads on a local sequence graph constructed for each targeted SV. This approach is different from other proposed and most existing graph methods that create a single whole-genome graph and align all reads to this large graph 18,39 . A whole-genome graph may be able to rescue reads from novel insertions that are misaligned to other parts of the genome in the original linear reference, however, the computational cost of building such a graph and performing alignment against this graph is very high.…”

“…Most targeted methods for genotyping are integrated with particular discovery algorithms and require the input SVs to be originally discovered by the designated SV caller [15][16][17] , require a complete genome-wide realignment 18,19 or need to be optimized on a set of training samples 12,20 .…”