Exploring Variation in the d N /d S Ratio Among Sites and Lineages Using Mutational Mappings: Applications to the Influenza Virus

Predicting Egg Passage Adaptations to Design Better Vaccines for the H3N2 Influenza Virus

Seasonal H3N2 influenza evolves rapidly, leading to an extremely poor vaccine efficacy. Substitutions employed during vaccine production using embryonated eggs (i.e., egg passage adaptation) contribute to the poor vaccine efficacy (VE), but the evolutionary mechanism remains elusive. Using an unprecedented number of hemagglutinin sequences (n = 89,853), we found that the fitness landscape of passage adaptation is dominated by pervasive epistasis between two leading residues (186 and 194) and multiple other positions. Convergent evolutionary paths driven by strong epistasis explain most of the variation in VE, which has resulted in extremely poor vaccines for the past decade. Leveraging the unique fitness landscape, we developed a novel machine learning model that can predict egg passage substitutions for any candidate vaccine strain before the passage experiment, providing a unique opportunity for the selection of optimal vaccine viruses. Our study presents one of the most comprehensive characterizations of the fitness landscape of a virus and demonstrates that evolutionary trajectories can be harnessed for improved influenza vaccines.

MADE: A Computational Tool for Predicting Vaccine Effectiveness for the Influenza A(H3N2) Virus Adapted to Embryonated Eggs

Seasonal Influenza H3N2 virus poses a great threat to public health, but its vaccine efficacy remains suboptimal. One critical step in influenza vaccine production is the viral passage in embryonated eggs. Recently, the strength of egg passage adaptation was found to be rapidly increasing with time driven by convergent evolution at a set of functionally important codons in the hemagglutinin (HA1). In this study, we aim to take advantage of the negative correlation between egg passage adaptation and vaccine effectiveness (VE) and develop a computational tool for selecting the best candidate vaccine virus (CVV) for vaccine production. Using a probabilistic approach known as mutational mapping, we characterized the pattern of sequence evolution driven by egg passage adaptation and developed a new metric known as the adaptive distance (AD) which measures the overall strength of egg passage adaptation. We found that AD is negatively correlated with the influenza H3N2 vaccine effectiveness (VE) and ~75% of the variability in VE can be explained by AD. Based on these findings, we developed a computational package that can Measure the Adaptive Distance and predict vaccine Effectiveness (MADE). MADE provides a powerful tool for the community to calibrate the effect of egg passage adaptation and select more reliable strains with minimum egg-passaged changes as the seasonal A/H3N2 influenza vaccine.

Passage Adaptation Correlates With the Reduced Efficacy of the Influenza Vaccine

BACKGROUND

As a dominant seasonal influenza virus, H3N2 virus rapidly evolves in humans and is a constant threat to public health. Despite sustained research efforts, the efficacy of H3N2 vaccine has decreased rapidly. Even though antigenic drift and passage adaptation (substitutions accumulated during vaccine production in embryonated eggs) have been implicated in reduced vaccine efficacy (VE), their respective contributions to the phenomenon remain controversial.

METHODS

We utilized mutational mapping, a powerful probabilistic method for studying sequence evolution, to analyze patterns of substitutions in different passage conditions for an unprecedented amount of H3N2 hemagglutinin sequences (n = 32 278).

RESULTS

We found that passage adaptation in embryonated eggs is driven by repeated convergent evolution over 12 codons. Based on substitution patterns at these sites, we developed a metric, adaptive distance (AD), to quantify the strength of passage adaptation and subsequently identified a strong negative correlation between AD and VE.

CONCLUSIONS

The high correlation between AD and VE implies that passage adaptation in embryonated eggs may be a strong contributor to the recent reduction in H3N2 VE. We developed a computational package called MADE (Measuring Adaptive Distance and vaccine Efficacy based on allelic barcodes) to measure the strength of passage adaptation and predict the efficacy of a candidate vaccine strain. Our findings shed light on strategies for reducing Darwinian evolution within the passaging medium in order to potentially restore an effective vaccine program in the future.

Calculating Higher-Order Moments of Phylogenetic Stochastic Mapping Summaries in Linear Time

Stochastic mapping is a simulation-based method for probabilistically mapping substitution histories onto phylogenies according to continuous-time Markov models of evolution. This technique can be used to infer properties of the evolutionary process on the phylogeny and, unlike parsimony-based mapping, conditions on the observed data to randomly draw substitution mappings that do not necessarily require the minimum number of events on a tree. Most stochastic mapping applications simulate substitution mappings only to estimate the mean and/or variance of two commonly used mapping summaries: the number of particular types of substitutions (labeled substitution counts) and the time spent in a particular group of states (labeled dwelling times) on the tree. Fast, simulation-free algorithms for calculating the mean of stochastic mapping summaries exist. Importantly, these algorithms scale linearly in the number of tips/leaves of the phylogenetic tree. However, to our knowledge, no such algorithm exists for calculating higher-order moments of stochastic mapping summaries. We present one such simulation-free dynamic programming algorithm that calculates prior and posterior mapping variances and scales linearly in the number of phylogeny tips. Our procedure suggests a general framework that can be used to efficiently compute higher-order moments of stochastic mapping summaries without simulations. We demonstrate the usefulness of our algorithm by extending previously developed statistical tests for rate variation across sites and for detecting evolutionarily conserved regions in genomic sequences.

Dynamic Convergent Evolution Drives the Passage Adaptation across 48 Years' History of H3N2 Influenza Evolution

Influenza viruses are often propagated in a diverse set of culturing media and additional substitutions known as passage adaptation can cause extra evolution in the target strain, leading to ineffective vaccines. Using 25,482 H3N2 HA1 sequences curated from Global Initiative on Sharing All Influenza Data and National Center for Biotechnology Information databases, we found that passage adaptation is a very dynamic process that changes over time and evolves in a seesaw like pattern. After crossing the species boundary from bird to human in 1968, the influenza H3N2 virus evolves to be better adapted to the human environment and passaging them in embryonated eggs (i.e., an avian environment) leads to increasingly stronger positive selection. On the contrary, passage adaptation to the mammalian cell lines changes from positive selection to negative selection. Using two statistical tests, we identified 19 codon positions around the receptor binding domain strongly contributing to passage adaptation in the embryonated egg. These sites show strong convergent evolution and overlap extensively with positively selected sites identified in humans, suggesting that passage adaptation can confound many of the earlier studies on influenza evolution. Interestingly, passage adaptation in recent years seems to target a few codon positions in antigenic surface epitopes, which makes it difficult to produce antigenically unaltered vaccines using embryonic eggs. Our study outlines another interesting scenario whereby both convergent and adaptive evolution are working in synchrony driving viral adaptation. Future studies from sequence analysis to vaccine production need to take careful consideration of passage adaptation.

A counting renaissance: combining stochastic mapping and empirical Bayes to quickly detect amino acid sites under positive selection

MOTIVATION

Statistical methods for comparing relative rates of synonymous and non-synonymous substitutions maintain a central role in detecting positive selection. To identify selection, researchers often estimate the ratio of these relative rates (dN/dS) at individual alignment sites. Fitting a codon substitution model that captures heterogeneity in dN/dS across sites provides a reliable way to perform such estimation, but it remains computationally prohibitive for massive datasets. By using crude estimates of the numbers of synonymous and non-synonymous substitutions at each site, counting approaches scale well to large datasets, but they fail to account for ancestral state reconstruction uncertainty and to provide site-specific dN/dS estimates.

RESULTS

We propose a hybrid solution that borrows the computational strength of counting methods, but augments these methods with empirical Bayes modeling to produce a relatively fast and reliable method capable of estimating site-specific dN/dS values in large datasets. Importantly, our hybrid approach, set in a Bayesian framework, integrates over the posterior distribution of phylogenies and ancestral reconstructions to quantify uncertainty about site-specific dN/dS estimates. Simulations demonstrate that this method competes well with more-principled statistical procedures and, in some cases, even outperforms them. We illustrate the utility of our method using human immunodeficiency virus, feline panleukopenia and canine parvovirus evolution examples.

Fast and robust characterization of time-heterogeneous sequence evolutionary processes using substitution mapping

Genes and genomes do not evolve similarly in all branches of the tree of life. Detecting and characterizing the heterogeneity in time, and between lineages, of the nucleotide (or amino acid) substitution process is an important goal of current molecular evolutionary research. This task is typically achieved through the use of non-homogeneous models of sequence evolution, which being highly parametrized and computationally-demanding are not appropriate for large-scale analyses. Here we investigate an alternative methodological option based on probabilistic substitution mapping. The idea is to first reconstruct the substitutional history of each site of an alignment under a homogeneous model of sequence evolution, then to characterize variations in the substitution process across lineages based on substitution counts. Using simulated and published datasets, we demonstrate that probabilistic substitution mapping is robust in that it typically provides accurate reconstruction of sequence ancestry even when the true process is heterogeneous, but a homogeneous model is adopted. Consequently, we show that the new approach is essentially as efficient as and extremely faster than (up to 25 000 times) existing methods, thus paving the way for a systematic survey of substitution process heterogeneity across genes and lineages.

Investigating protein-coding sequence evolution with probabilistic codon substitution models

This review is motivated by the true explosion in the number of recent studies both developing and ameliorating probabilistic models of codon evolution. Traditionally parametric, the first codon models focused on estimating the effects of selective pressure on the protein via an explicit parameter in the maximum likelihood framework. Likelihood ratio tests of nested codon models armed the biologists with powerful tools, which provided unambiguous evidence for positive selection in real data. This, in turn, triggered a new wave of methodological developments. The new generation of models views the codon evolution process in a more sophisticated way, relaxing several mathematical assumptions. These models make a greater use of physicochemical amino acid properties, genetic code machinery, and the large amounts of data from the public domain. The overview of the most recent advances on modeling codon evolution is presented here, and a wide range of their applications to real data is discussed. On the downside, availability of a large variety of models, each accounting for various biological factors, increases the margin for misinterpretation; the biological meaning of certain parameters may vary among models, and model selection procedures also deserve greater attention. Solid understanding of the modeling assumptions and their applicability is essential for successful statistical data analysis.

The quest for natural selection in the age of comparative genomics

Continued genome sequencing has fueled progress in statistical methods for understanding the action of natural selection at the molecular level. This article reviews various statistical techniques (and their applicability) for detecting adaptation events and the functional divergence of proteins. As large-scale automated studies become more frequent, they provide a useful resource for generating biological null hypotheses for further experimental and statistical testing. Furthermore, they shed light on typical patterns of lineage-specific evolution of organisms, on the functional and structural evolution of protein families and on the interplay between the two. More complex models are being developed to better reflect the underlying biological and chemical processes and to complement simpler statistical models. Linking molecular processes to their statistical signatures in genomes can be demanding, and the proper application of statistical models is discussed.