Session V66: Microbial and Viral Quantitative Evolution

RNA virus evolution and the predictability of next year's flu

RNA viruses like HIV of influenza virus evolve rapidly and thereby evade human immunity. We have performed whole genome deep sequencing of longitudinal samples spanning 5-10 years from acute to chronic HIV infection. From these data, we estimate the landscape of fitness costs at every nucleotide in the genome and characterize the the complex dynamics of immune escape and reversion. Similarly, influenza viruses change their antigenic properties rapidly and the seasonal influenza vaccine needs to be updated whenever the viruses change antigenically. We show how sequence data can be used predict which influenza virus variants are most likely to succeed and circulate in future seasons. These predictions use theoretical insights into the structure of genealogies of rapidly adapting populations and are provided in near real-time at nextflu.org.


  

Single nucleotide mapping of locally accessible trait space in evolving yeast

Tradeoffs constrain the improvement of performance of all traits simultaneously. Such tradeoffs define a Pareto optimality front that represents a set of optimal individuals that cannot be improved in any one trait without reducing performance in another trait. While widely assumed, direct experimental evolutionary approaches often fail to detect tradeoffs with many experiments generating individuals that improve performance in all measured traits. Moreover, even when an improvement in one trait is found to be associated with the loss of performance in another, it is hard to establish that such a negative correlation is not induced by the specific features of sampled mutations and that other possible adaptive mutations cannot escape such apparent tradeoffs. Here we detect tradeoffs and define the Pareto optimality front in the context of short-term evolution of S. cerevisiae in glucose-limited media. We have evolved barcoded yeast populations under several conditions, with each condition selecting for improved performance in different parts of the yeast growth cycle. By isolating hundreds of adaptive clones and quantifying their performances in each growth part of the cycle, we defined tradeoffs between performances in fermentation and respiration and respiration and stationary phase. Importantly, due to the large numbers of the studied clones we were able to claim that no single point mutation in the yeast genome can improve the performance beyond either of the defined optimality fronts. We found that in both cases the shape of the optimality front is convex suggesting the possibility of short term evolution to select for generalists. Finally, by sequencing hundreds of adaptive clones, we identified the molecular basis underlying identified trade-offs and revealed novel targets of adaptation.   

Tracking pathogenetic bacteria at the plasmid, genome, and pan-genomic level.

The rapid global increase of multidrug-resistant organisms presents a major global health threat that will dramatically reduce the efficacy of antibiotics and thus constrain the number of effective treatments available to patients. As opposed to analogous efforts in viral epidemiology, accurate reconstruction of the pandemic spread of antibiotic resistance remains intractable for reasonable sample sizes due, in large part, to the high rate of homologous recombination and horizontal gene transfer that prevents the application of traditional phylogenetic approaches. Lastly, complete assemblies are a prerequisite to such quantitative study.

In this talk, I will present a novel computational framework for bacterial evolution that generalizes the traditional linear reference genome to a pan-genomic non-planar graph. Using this framework, I will analyze 120 antibiotic resistant genomes collected in Basel, Switzerland over the course of a decade, sequenced with both ONT and Illumina data. We show that the evolution of antibiotic resistance exhibits a nested doll structure in which genetic transposition, homologous recombination, and clonal expansion occur at similar time-scales.   

Investigating the role of hemagglutinin protein stability in influenza A evolutionary dynamics

Influenza A viruses (IAVs) cause a significant burden to human populations. They are notable for their rapid evolution arising from error-prone RNA polymerases in conjunction with selective pressures from hosts. Most host immune responses target the surface protein hemagglutinin (HA) and typically generate lifelong antibodies to seen strains. Thus to survive, IAVs must undergo mutations in HA that allow immune escape, while the HA protein must also satisfy functional constraints. We formulate a mathematical model of IAV evolution and transmission that spans scales from molecular to population processes. We couple models for HA protein thermal stability, mutation, and cross-immunity with a population-level model for the spread of infection, and characterize the effects of stability. Certain residues buried in the core of the HA protein may affect stability but not immunogenicity. We adapt our model to include these “buried” residues, and evaluate their influence on viral diversity, transmission, and evolutionary dynamics.   

Dynamics of lineage diversity during adaptation to weak antibiotic pressure

Tracking low-frequency lineages in large microbial populations is key to fully elucidating their evolutionary dynamics, especially under weak selection pressures and short time scales. However, resolution is often limited by methods for labeling lineages or whole-genome sequencing. To overcome this challenge, we introduce a large library of random DNA barcodes into E. coli, allowing us to track lineages down to tens or hundreds of cells. We observe two distinct phases of lineage dynamics during adaptation to low levels of antibiotics. During the first phase, the diversity of lineages undergoes a rapid and predictable drop which is characteristic of the drug and concentration. In particular, we find that low amounts of trimethoprim actually slow down the loss of diversity compared to the absence of drug. This initial loss of diversity appears to be driven by selection on standing genetic variation, leading to groups of lineages rising or falling together both within and between populations. During the second phase of dynamics, new mutations arise on these lineages, leading to clonal interference. Lineage diversity then stabilizes at a low but nonzero level, which appears to be universal across conditions despite the variation in initial dynamics.   

Testing the Retroelement Invasion Hypothesis for the Emergence of the Ancestral Eukaryotic Cell

Phylogenetic evidence suggests that the ancestral eukaryotic cell emerged as a result of invasion and proliferation of retroelements, selfish mobile genetic elements that copy and paste themselves within a host genome. Here we test this hypothesis by determining the pressures retroelements exert on simple genomes. We transferred two retroelements, human LINE-1 and the bacterial group II intron Ll.LtrB, into bacteria, and find that both are functional and detrimental to growth. We find, surprisingly, that retroelement lethality and proliferation is enhanced by the ability to perform eukaryotic-like nonhomologous end-joining (NHEJ) DNA repair. We show that the only stable evolutionary consequence in simple cells is maintenance of retroelements in low numbers, and that retrotransposition in eukaryotes must be finely tuned to allow proliferation.   

Physical Constraints on Epistatic Interactions

Epistasis, or the context-dependence of mutations, makes evolutionary trajectories difficult to predict. Here, we argue that knowledge of the physics of a system can tame its apparent evolutionary complexity. We focus on proteins, inspired by recent work on low-energy mechanical modes of bio-inspired, functional mechanical networks. Mutations in the sequence of a protein affect its shape, thereby modulating the effect of subsequent mutations; we find epistatic interactions at all probed orders. However, the structure of deformations limits the number of independent epistatic parameters. We find that such constrained epistasis is reflected in the topology of the fitness landscape. In addition to providing a mechanistic basis for experimentally observed epistatic couplings in proteins, we demonstrate how similar considerations may constrain epistasis in other complex systems.   

Effects of clonal interference on adaptation in a fixed fitness landscape

Adaptation in a fixed environment, where as a population evolves its fitness increases, is typically thought of as a hill climbing process
on a fitness landscape. On a fitness-parameterized landscape, the fitness trajectory depends on the statistics of mutation effects and
how they vary with current fitness ('macroscopic epistasis'). Results on fitness-parameterized landscapes are typically obtained in the
strong selection, weak mutation regime, where mutations fix or go extinct sequentially. In a large population, while the phenomenon of clonal interference is known to slow down the rate of evolution, it is unclear how the shape of the fitness trajectory is affected. Here, we
investigate the effects of clonal interference on fitness evolution on various fitness landscapes, including ones that explicitly takes into
account microscopic epistasis between mutation sites.   

Disentangling effects of genetic linkage on estimates of selection in intrahost HIV evolution

The evolutionary history of a population contains information about the underlying forces driving population diversity and adaptation. However, it is difficult for current methods to disentangle the effects of individual mutations from complex evolutionary dynamics. Here we describe a method to infer selection from genetic time-series data while also accounting for the confounding effects of genetic linkage, using a path integral approach based in statistical physics. We apply this method to investigate within-host HIV-1 evolution at the half-genome scale. Our approach reveals selective pressures for escape from human immune responses. Due to clonal interference between escape mutants, we find that accounting for genetic linkage is crucial to infer realistic estimates of selection. Our ability to account for genetic linkage is also critical for insight into complex evolutionary scenarios, highlighted by a struggle for dominance between co-infecting strains of the virus in an individual who ultimately develops broadly neutralizing antibodies.   

Dynamics of within-host evolution and strain replacement in the human gut microbiome

The human gut contains trillions of rapidly reproducing bacteria, making it a potential hotbed of within-host evolution. While the species-level dynamics in the gut have been extensively studied, we currently know very little about the evolutionary dynamics that take place within individual species. In this talk, I will discuss our recent efforts to measure the statistical properties of within-host evolution from a large panel of human gut metagenomes. Our analysis shows that on short timescales, genetic differences are only rarely caused by the invasion of distantly related strains. Resident strains more commonly acquire putative evolutionary changes, in which small numbers of nucleotide variants or gene content differences rapidly sweep to high frequency within a host. However, comparisons of adult twins suggest that strain replacement eventually overwhelms evolution over multi-decade timescales, hinting at fundamental limits to the extent of local adaptation. Together, our results show that gut bacteria can evolve on human-relevant timescales, and they provide key empirical constraints necessary for future modeling efforts.
  

Tuning Evolution Towards Generalists Through Resonant Environmental Cycling

Natural environments can present diverse fitness pressures, but some genotypes remain fit across a wide range of challenges. Such 'generalist' genotypes can be hard to evolve because there may be entropic or absolute fitness costs relative to specialist genotypes. Here, we study the conditions under which time-dependent evolutionary protocols stabilize generalists even when static protocols fail. We find that cycling environments on timescales tuned to match fixation times can reliably evolve generalists when the landscape is too rugged, or deleterious selection too adverse, for static protocols to succeed. We discuss 'chirp' protocols that circumvent the need for protocol fine-tuning. Our work reveals regimes in which time-dependent 'seascapes’ can find and stabilize populations around genotypes that are fundamentally unstable in any static protocol.