Novel and extendable genotyping system for human respiratory syncytial virus based on whole-genome sequence analysis

Abstract

Background: Human respiratory syncytial virus (RSV) is one of the leading causes of respiratory infections, especially in infants and young children. Previous RSV sequencing studies have primarily focused on partial sequencing of G gene (200-300 nucleotides) for genotype characterization or diagnostics. However, the genotype assignment with G gene has not recapitulated the phylogenetic signal of other genes, and there is no consensus on RSV genotype definition.

Methods: We conducted maximum likelihood phylogenetic analysis with 10 RSV individual genes and whole-genome sequence (WGS) that are published in GenBank. RSV genotypes were determined by using phylogenetic analysis and pair-wise node distances.

Results: In this study, we first statistically examined the phylogenetic incongruence, rate variation for each RSV gene sequence and WGS. We then proposed a new RSV genotyping system based on a comparative analysis of WGS and the temporal distribution of strains. We also provide an RSV classification tool to perform RSV genotype assignment and a publicly accessible up-to-date instance of Nextstrain where the phylogenetic relationship of all genotypes can be explored.

Conclusions: This revised RSV genotyping system will provide important information for disease surveillance, epidemiology, and vaccine development.

Keywords: genotype; genotypic classification; label software; phylogenetic analysis.

FIGURE 1

Scheme of respiratory syncytial virus (RSV) genome and comparison of RSV phylogenies inferred from different gene datasets. (A) RSV genome organization with G gene indels. (B) Likelihood scores of phylogenies inferred from different gene sequences given to the whole‐genome sequence (WGS) dataset. W1, the phylogeny inferred from WGS with G gene indels implemented as a single evolutionary event; W2, the phylogeny inferred from WGS with G gene indels implemented as multiple substitution events. *P < 0.005 in Shimodaira–Hasegawa (SH) test compared with W1; ^# P < 0.005 in approximately unbiased (AU) test compared with W1. (C) Comparison of evolutionary rates that are estimated from different gene datasets. Error bars indicate the confidence intervals of the estimation

FIGURE 2

Maximum likelihood phylogeny of respiratory syncytial virus (RSV) RSV‐A (A) and RSV‐B (B) inferred from whole‐genome sequence (WGS) with the genotyping assignment. The genotype assignments are indicated with vertical black bars and are labeled on the right. The subclade assignments are indicated with vertical gray bars and are labeled on the left. Tip point colors represent the previously defined genotype names based on complete or partial G gene sequences. The nodes that define the genotype and subclade are indicated with black and gray node points, respectively. Bootstrap of each ancestral genotype/subclade node is marked. Colored columns on the right side represent G gene duplication and indels. Scale bars indicate 0.01 nucleotide substitution per site

FIGURE 3

Spatial and temporal distribution of respiratory syncytial virus (RSV) RSV‐A (A) and RSV‐B (B) genotypes. The temporal and spatial distribution of RSV genotypes is based on the detection year and isolated WHO region of sequence for each assigned genotype. African Region (AFRO), Region of the Americas (PAHO), South‐East Asia Region (SEARO), European Region (EURO), Eastern Mediterranean Region (EMRO), and Western Pacific Region (WPRO)