Predicting the Mutating Distribution at Antigenic Sites of the Influenza Virus.

Hongyang Xu,Qingchen Zhang,Tianyi Qiu,Li Jin,Kailin Tang,Yiyan Yang,Jingxuan Qiu,Shuning Wang,Yungang He,Ruixin Zhu,Zhiwei Cao

doi:10.1038/srep20239

Abstract

Mutations of the influenza virus lead to antigenic changes that cause recurrent epidemics and vaccine resistance. Preventive measures would benefit greatly from the ability to predict the potential distribution of new antigenic sites in future strains. By leveraging the extensive historical records of HA sequences for 90 years, we designed a computational model to simulate the dynamic evolution of antigenic sites in A/H1N1. With templates of antigenic sequences, the model can effectively predict the potential distribution of future antigenic mutants. Validation on 10932 HA sequences from the last 16 years showing that the mutated antigenic sites of over 94% of reported strains fell in our predicted profile. Meanwhile, our model can successfully capture 96% of antigenic sites in those dominant epitopes. Similar results are observed on the complete set of H3N2 historical data, supporting the general applicability of our model to multiple sub-types of influenza. Our results suggest that the mutational profile of future antigenic sites can be predicted based on historical evolutionary traces despite the widespread, random mutations in influenza. Coupled with closely monitored sequence data from influenza surveillance networks, our method can help to forecast changes in viral antigenicity for seasonal flu and inform public health interventions.

Highlights

The seasonal influenza virus is well known for its rapid mutation rate and constant antigenic changes, which causes a major and persistent challenge to public health
The assumptions of our model include the following: 1) the major antigen of HA experiences greater evolutional pressure compared with other proteins in the virus, and deserve an antigen-specific evolutional model instead of a model at genome level; 2) residual diversities at different positions of antigenic sites often imply different adaptive abilities, such as contact transmissibility[10] and immune-escaping ability, while the adaptive ability might be partially inferred from the historical trace of antigenic positions; and 3) the final dominance of an antigenic mutant may be related to its inherent adaptive ability, and its population
A computational model was designed to simulate the future distribution of new antigenic mutants based on the evolutionary footprints at antigenic sites

Summary

Introduction

The seasonal influenza virus is well known for its rapid mutation rate and constant antigenic changes, which causes a major and persistent challenge to public health. To better understand these changes, previous studies have established evolutionary models to trace back the genomic variations and epidemiological dynamics at genome level[1,2]. The rapid accumulation of high-quality genome sequences has provided new opportunities to analyse virus spread and phylodynamics based on current and past influenza genomes aiming for better preventive measures[3,4]. When the future distribution of new epitope mutants is obtained for major antigens, can better preventive measures be achieved. Our results revealed that the future mutant profiles of the HA antigen sites are predictable, including those dominant antigenic sequences

Methods

Results

Conclusion