Detection of Emerging Vaccine-Related Polioviruses by Deep Sequencing

Figures

• 

  FIG 1. Determination of the error threshold. (A) Error events observed across the genome in low-passage-number controls. Insertions (ins), deletions (del), and substitutions (subst) are color coded. Note the short error-prone region (bp 810 to 900) in Sabin 2. The solid black line represents the overall per base error rate, 0.0033. (B) Error events plotted as a function of read depth. (C) Prediction of insertion, deletion, and substitution error rates based on sequence depth using a general linear model (Poisson) with observed error data. (D) Comparison of overall errors using Poisson and negative-binomial (gamma-Poisson mixture) models. The black curve represents the error threshold utilized in this study. The threshold was calculated from the overall per base error rate using the Poisson distribution at a P of <0.001.

• 

  FIG 2. Variant analysis of vaccine-related polioviruses. (A) Distribution of synonymous and nonsynonymous variants in the 5′ UTR and VP region of Sabin 1, Sabin 2, and Sabin 3 vaccine-related polioviruses. Horizontal black bars, medians. (B) Distribution of variant abundance in the 5′ UTR and VP region. Horizontal black bars, medians.

• 

  FIG 3. Frequency of VP1 nucleotide variants at each amino acid position in Sabin 1, Sabin 2, and Sabin 3. Dotted lines mark amino acid positions associated with phenotypic reversion. Variants include nonsynonymous and synonymous changes. AA, amino acid.