EXPLORING VIRUS SEQUENCE DIVERSITY USING GENOME GRAPHS

Abstract

Linear reference genomes are a consensus of the most frequent base at a given position. Graph-based reference genomes represent a genome as a network of alternative paths, thereby providing information on variant bases. This graph can be node labelled—bases are held in the nodes of the graph while the edges represent the connections between the bases, or edge labelled—if the edges in the graph represent the bases. A graph-based data structure like this is suitable for exploring and describing virus sequence diversity. Sequenced respiratory syncytial virus raw reads from a twenty-five-member household collected during a household outbreak were used to generate a genome graph. The sample reads were then aligned to the genome graph. The number of reads that mapped under each node was then summarised as a multidimensional coverage vector. A pairwise Euclidean distance matrix was computed and a neighbour-joining cladogram based on hierarchical clustering generated. We demonstrate the plausibility of differentiating a large number of closely related consensus genomes by comparing the number of respective raw reads that align to each node from the larger genome graph. Additionally, the sequence coverage across a genome graph provides an alternative approach for examining sequence relatedness and identifying potential sequencing errors.