Assessment of the Utility of Whole Genome Sequencing of Measles Virus in the Characterisation of Outbreaks

Abstract

Background: Measles is a highly infectious disease caused by measles virus (MeV). Despite the availability of a safe and cost-effective vaccine, measles is one of the world-leading causes of death in young children. Within Europe, there is a target for eliminating endemic measles in 2015, with molecular epidemiology required on 80% of cases for inclusion/exclusion of outbreak transmission chains. Currently, MeV is genotyped on the basis of a 450 nucleotide region of the nucleoprotein gene (N-450) and the hemagglutinin gene (H). However, this is not sufficiently informative for distinguishing endemic from imported MeV. We have developed an amplicon-based method for obtaining whole genome sequences (WGS) using NGS or Sanger methodologies from cell culture isolates or oral fluid specimens, and have sequenced over 60 samples, including 42 from the 2012 outbreak in the UK.

Results: Overall, NGS coverage was over 90% for approximately 71% of the samples tested. Analysis of 32 WGS excluding 3' and 5' termini (WGS-t) obtained from the outbreak indicates that the single nucleotide difference found between the two major groups of N-450 sequences detected during the outbreak is most likely a result of stochastic viral mutation during endemic transmission rather than of multiple importation events: earlier strains appear to have evolved into two distinct strain clusters in 2013, one containing strains with both outbreak-associated N-450 sequences. Additionally, phylogenetic analysis of each genomic region of MeV for the strains in this study suggests that the most information is acquired from the non-coding region located between the matrix and fusion protein genes (M/F NCR) and the N-450 genotyping sequence, an observation supported by entropy analysis across genotypes.

Conclusions: We suggest that both M/F NCR and WGS-t could be used to complement the information from classical epidemiology and N-450 sequencing to address specific questions in the context of measles elimination.

Fig 1. Schematic representation of the measles virus genome.

The 15,894 nucleotides (nt) of the measles virus genome encode: nucleoprotein (N; 525 aa), phosphoprotein (P; 507 aa), matrix (M; 335 aa), fusion (F; 550 aa), hemagglutinin (H; 617 aa), large polymerase (L; 2,183 aa), C (299 aa) and V (186 aa) proteins. Coding regions of the genome (in white) are separated by non-coding regions (NCR; in grey). The longest NCR is that between the M and F genes: M/F NCR (1,012 nt). (aa used as abbreviation for amino acid)

Fig 2. Sequencing outcomes.

(A) Percentage of samples for which the coverage obtained falls within each 5% coverage window following sequencing by NGS and Sanger (i) or NGS only (ii). Bars exclude lower category value and include higher category value (except first bar, which includes 0). (B) The fraction of the genome obtained (% coverage) by a combination of NGS and Sanger (black diamonds) or NGS only (grey squares) is plotted against sample viral titre (genome copies / μl).

Fig 3. Phylogenetic analysis of the N-450, H, N and M/F NCR sequences of D8 outbreak strains.

The 32 D8 strains for which the whole genome sequence excluding the 3’ and 5’ termini (WGS-t) was available were analysed using RAxML. Phylogenetic trees were obtained for the N-450 (A), H (B), N (C) and M/F NCR (D) sequences. Strains are coloured according to the epidemiology clusters in which they were grouped.

Fig 4. Phylogenetic analysis of the WGS-t sequence of D8 outbreak strains.

Phylogenetic analysis was carried out as described for Fig 3. A smaller scale is used for the WGS-t phylogenetic tree given the lower rate of changes/site.

Fig 5. BEAST analysis of D8 strains.

Analysis of the M/F NCR (A) and WGS-t (B) sequences of the D8 strains in this study was carried out using the week and year of sample collection as a measurement of sample time.

Fig 6. Sequence variability in the measles virus genome.

Shandon entropy plot representing nucleotide variability at each position of the measles genome (average over a sliding window of 100 nucleotides with a step of 50 bases). Analysis was carried out for D8 (black line) and all available measles genome sequences in GenBank for measles genome sequences from non-A genotype (grey line).