Using Statistical Phylogenetics for Investigation of Enterovirus 71 Genotype A Reintroduction into Circulation

Abstract

Neurovirulent enterovirus 71 (EV-A71) caused a massive epidemic in China in 2008–2011. While subgenotype C4 was the major causative agent, a few isolates were almost identical to the prototype EV-A71 strain and belonged to genotype A. This variant was allegedly extinct since 1970, and its identification in this epidemic suggests reintroduction of the archive virus. Regression analysis of genetic distances (TempEst software) was of moderate utility due to the low resolution of classical phylogenetic methods. Bayesian phylogenetic analysis (BEAST software) suggested artificial introduction event based on highly aberrant phylogenetic tree branch rates that differed by over three standard deviations from the mean substitution rate for EV71. Manual nucleotide-level analysis was used to further explore the virus spread pattern after introduction into circulation. Upon reintroduction, the virus accumulated up to seven substitutions in VP1, most of them non-synonymous and located within the capsid’s canyon or at its rims, compatible with readaptation of a lab strain to natural circulation.

Molecular epidemiology and recombination of Enterovirus A71 in mainland China from 1987 to 2017

“…Based on phylogenetic analysis of VP1 gene, EV-A71 is currently classified into 8 genotypes A–H (Deshpande et al, 2003 ; Saxena et al, 2015 ; Bessaud et al, 2014 ; Majumdar et al, 2018 ). Genotype A contained a sole prototype strain BrCr isolated in California, USA, in 1969 (Schmidt et al, 1974 ) and some strains emerged in China in 2008–2010 (Vakulenko et al, 2019 ). The emergence of genotype A EV-A71 strains in China was considered as a release of the prototype strain from a laboratory into circulation (Vakulenko et al, 2019 ).…”

“…Genotype A contained a sole prototype strain BrCr isolated in California, USA, in 1969 (Schmidt et al, 1974 ) and some strains emerged in China in 2008–2010 (Vakulenko et al, 2019 ). The emergence of genotype A EV-A71 strains in China was considered as a release of the prototype strain from a laboratory into circulation (Vakulenko et al, 2019 ). Genotypes B and C are subdivided into subgenotypes B0–B5 and C0–C5, respectively (Noisumdaeng et al, 2018 ).…”

The Effect of Sample Bias and Experimental Artefacts on the Statistical Phylogenetic Analysis of Picornaviruses

“…As entries with a supposedly incorrect isolation date annotation are not uncommon in GenBank, correlation analysis and verification of the outlier sequences is strongly recommended prior to performing a statistical phylogenetic study. It is noteworthy that the correlation analysis can be much less sensitive on more heterogeneous datasets, such as the whole EV-A71 type dataset [55].…”

“…Regression analysis of isolation time and root-to-tip distance allows for the identification of date annotation errors and is necessary in the case of enteroviruses and other picornaviruses. The distribution of branch rates (or their logarithms) of the MCC tree should be close to normal for most picornaviruses, due to their biological properties, and is a good way to detect unnatural events, such as annotation errors or the release of an archive virus [55]. The criteria for the exclusion of sequences remain arbitrary and should consider many factors, especially because certain natural events, such as rapid re-adaptation of viruses to a new system [56] or long-term persistence [57], can result in aberrant rates in a virus lineage.…”

Special Issue “Human Picornaviruses”

“…Papers about virus recombination and recombinants identified by means of next-generation sequencing and bioinformatics, viral cell surface receptors, virus replication, interaction with host cells, and immune responses are included. While several papers (Osundare et al [1], Chamings et al [2], Vakulenko et al [3] and Hietanen & Susi [4]) deal with generic amplification, sequencing, identification and/or typing of viral sequences and viruses, the second paper by Vakulenko et al [5] is of major importance, as it deals with the effects of sample bias and experimental artefacts on the outcome of phylogenetic analyses. The results suggest that even a single erroneous sequence may profoundly destabilize the whole analysis by increasing the variance of the inferred evolutionary parameters.…”