Sequence Alignment on Directed Graphs

Abstract

Genomic variations in a reference collection are naturally represented as genome variation graphs. Such graphs encode common subsequences as vertices and the variations are captured using additional vertices and directed edges. The resulting graphs are directed graphs possibly with cycles. Existing algorithms for aligning sequences on such graphs make use of partial order alignment (POA) techniques that work on directed acyclic graphs (DAGs). To achieve this, acyclic extensions of the input graphs are first constructed through expensive loop unrolling steps (DAGification). Furthermore, such graph extensions could have considerable blowup in their size and in the worst case the blow-up factor is proportional to the input sequence length. We provide a novel alignment algorithm V-ALIGN that aligns the input sequence directly on the input graph while avoiding such expensive DAGification steps. V-ALIGN is based on a novel dynamic programming (DP) formulation that allows gapped alignment directly on the input graph. It supports affine and linear gaps. We also propose refinements to V-ALIGN for better performance in practice. With the proposed refinements, the time to fill the DP table has linear dependence on the sizes of the sequence, the graph, and its feedback vertex set. We conducted experiments to compare the proposed algorithm against the existing POA-based techniques. We also performed alignment experiments on the genome variation graphs constructed from the 1000 Genomes data. For aligning short sequences, standard approaches restrict the expensive gapped alignment to small filtered subgraphs having high similarity to the input sequence. In such cases, the performance of V-ALIGN for gapped alignment on the filtered subgraph depends on the subgraph sizes.

AStarix: Fast and Optimal Sequence-to-Graph Alignment

“…Many of these, however, are designed only for specific types of genome graphs, such as de Bruijn graphs [24,11,23] and variation graphs [9]. A compromise often made when aligning sequences to cyclic graphs using algorithms reliant on directed acyclic graphs involves the computationally expensive "DAG-ification" of graph regions [18,9].…”

“…, e k ) in G r induces a spelling σ(π) ∈ Σ * defined by σ(e 1 ) · · · σ(e k ), where σ(e i ) is the label of edge e i and Σ * := k∈N Σ k . We note that our approach naturally handles cyclic walks and does not require cycle unrolling, a feature shared with GraphAligner [27] and Brown-ieAligner [11] but missing from vg [9], PaSGAL [15] and V-ALIGN [18].…”

GraphAligner: Rapid and Versatile Sequence-to-Graph Alignment

“…Although POA is defined only for acyclic graphs, it can be extended to cyclic graphs by unfolding cyclic components, which is the approach taken by the VG toolkit [16] and ExpansionHunter [9]. The practical efficiency of this unfolding depends on the read length and the graph topology and complex cyclic areas can lead to very large unfolded graphs [20]. V-Align [20] aligns to cyclic graphs but its runtime depends on the graph's feedback vertex set size.…”

“…The practical efficiency of this unfolding depends on the read length and the graph topology and complex cyclic areas can lead to very large unfolded graphs [20]. V-Align [20] aligns to cyclic graphs but its runtime depends on the graph's feedback vertex set size. Some tools use heuristic approaches for aligning to de Bruijn graphs using depth-first search [6,21,8].…”

Third international conference “Bioinformatics: from Algorithms to Applications” (BiATA 2019)

“…Many popular short read assemblers [9,10,11] provide the user not only with a set of contig sequences, but also with assembly graphs, encoding the information on the potential adjacencies of the assembled sequences. Naturally arising problem of sequence-to-graph alignment has been a topic of many recent studies [2,3,5,6,7,8,13,14]. Identifying alignments of long error-prone reads (such as Pacbio and ONT reads) to assembly graphs is particularly important and has recently been applied to hybrid genome assembly [1,4], read error correction [12], and haplotype separation [3].…”