A41 Deep sequencing of respiratory syncytial virus links viral diversity to disease severity

Respiratory syncytial virus (RSV) is a common virus that can cause bronchiolitis in infants and pneumonia in immunocompromised and elderly people. RSV belongs to the Pneumoviridae family and consists of a genome of 15 kb. Its genome contains ten genes that code for eleven proteins, with M2 coding for two different proteins in overlapping open reading frames. It is unclear why some infected children have severe disease and others have mild or asymptomatic disease. In this project, methods for complete genome sequencing of RSV via Sanger and Illumina MiSeq platforms were optimized. One hundred and twenty-four community samples (59 RSV A and 65 RSV B) from 2014 to 2018 were collected (in collaboration with the Royal College of General Practitioners) and sequenced. Samples were selected based on viral load (e.g. Ct values had to be < 30). The genotype of each sample was determined by constructing phylogenetic trees with reference sequences from all genotypes. Trees were reconstructed using the maximum likelihood method. Furthermore, Illumina sequencing was used to deep sequence seven community samples and four hospital samples that were spatiotemporally matched (obtained via Imperial College NHS Trust hospitals). Variants were studied to investigate if certain variants influence disease severity (e.g. cause mild (community samples) or severe infection (hospital samples)). Analysis so far showed that ON1 (with a seventy-two nucleotide duplication in attachment protein G) is the most common genotype in both community and hospitalized samples (90% and 75% of samples, respectively), with GA2 (without duplication) as the next most common genotype for RSV A subtypes (7% and 25%). Three per cent of community samples were of the GA5 genotype. Samples from the RSV B subgroup all belong to the BA genotypes with a 60-nucleotide duplication in G. Samples that were selected for Illumina sequencing had a Ct value between 19.0 and 29.1, while hospital samples had a Ct value of 18.3 to 29.1. Viral load, therefore, did not explain disease severity in these selected samples. The Shannon entropy from Illumina sequenced samples averaged at 22.78 in community samples (ranges from 15 to 28) and 38.78 in hospitalized samples (ranges from 31 to 57). This indicated that diversity of the virus pool might influence disease severity; however, more samples need to be analyzed. There are no specific variants that could explain disease severity. Diversity of the virus pool could explain the link between higher viral loads and disease severity, which is sometimes found but cannot always be confirmed. Higher viral loads can harbor more diverse viral particles compared to lower viral loads. Future work will focus on more in-depth variation and diversity analysis and on evolutionary analysis of both community and hospital samples. We will also investigate intra-host evolution of RSV in acute infections using consecutive samples and its possible implications on the host response.