Self-replicating artificial neural networks give rise to universal evolutionary dynamics.

In evolutionary models, mutations are exogenously introduced by the modeler, rather than endogenously introduced by the replicator itself. We present a new deep-learning based computational model, the self-replicating artificial neural network (SeRANN). We train it to (i) copy its own genotype, like a biological organism, which introduces endogenous spontaneous mutations; and (ii) simultaneously perform a classification task that determines its fertility. Evolving 1,000 SeRANNs for 6,000 generations, we observed various evolutionary phenomena such as adaptation, clonal interference, epistasis, and evolution of both the mutation rate and the distribution of fitness effects of new mutations. Our results demonstrate that universal evolutionary phenomena can naturally emerge in a self-replicator model when both selection and mutation are implicit and endogenous. We therefore suggest that SeRANN can be applied to explore and test various evolutionary dynamics and hypotheses.

Aneuploidy Can Be an Evolutionary Diversion on the Path to Adaptation.

Aneuploidy is common in eukaryotes, often leading to decreased fitness. However, evidence from fungi and human tumur cells suggests that specific aneuploidies can be beneficial under stressful conditions and facilitate adaptation. In a previous evolutionary experiment with yeast, populations evolving under heat stress became aneuploid, only to later revert to euploidy after beneficial mutations accumulated. It was therefore suggested that aneuploidy is a "stepping stone" on the path to adaptation. Here, we test this hypothesis. We use Bayesian inference to fit an evolutionary model with both aneuploidy and mutation to the experimental results. We then predict the genotype frequency dynamics during the experiment, demonstrating that most of the evolved euploid population likely did not descend from aneuploid cells, but rather from the euploid wild-type population. Our model shows how the beneficial mutation supply-the product of population size and beneficial mutation rate-determines the evolutionary dynamics: with low supply, much of the evolved population descends from aneuploid cells; but with high supply, beneficial mutations are generated fast enough to outcompete aneuploidy due to its inherent fitness cost. Our results suggest that despite its potential fitness benefits under stress, aneuploidy can be an evolutionary "diversion" rather than a "stepping stone": it can delay, rather than facilitate, the adaptation of the population, and cells that become aneuploid may leave less descendants compared to cells that remain diploid.

Cultural transmission, competition for prey, and the evolution of cooperative hunting.

Although cooperative hunting is widespread among animals, its benefits are unclear. At low frequencies, cooperative hunting may allow predators to escape competition and access bigger prey that could not be caught by a lone cooperative predator. Cooperative hunting is a more successful strategy when it is common, but its spread can result in overhunting big prey, which may have a lower per-capita growth rate than small prey. We construct a one-predator species, two-prey species model in which predators either learn to hunt small prey alone or learn to hunt big prey cooperatively. Predators first learn vertically from parents, then horizontally (i.e. socially) from random individuals or siblings. After horizontal transmission, they hunt with their learning partner if both are cooperative, and otherwise they hunt alone. Cooperative hunting cannot evolve when initially rare unless predators (a) interact with siblings, or (b) horizontally transmit the cooperative behavior to potential hunting partners. Whereas competition for small prey favors cooperative hunting when this cooperation is initially rare, the frequency of cooperative hunting cannot reach 100% unless big prey is abundant. Furthermore, a mutant that increases horizontal learning can invade if cooperative hunting is present, but not at 100%, because horizontal learning allows pairs of predators to have the same strategy. Our results reveal that the interactions between prey availability, social learning, and degree of cooperation among predators may have important effects on ecosystems.

Phenotype switching of the mutation rate facilitates adaptive evolution.

The mutation rate plays an important role in adaptive evolution. It can be modified by mutator and anti-mutator alleles. Recent empirical evidence hints that the mutation rate may vary among genetically identical individuals: evidence from bacteria suggests that the mutation rate can be affected by expression noise of a DNA repair protein and potentially also by translation errors in various proteins. Importantly, this non-genetic variation may be heritable via a transgenerational epigenetic mode of inheritance, giving rise to a mutator phenotype that is independent from mutator alleles. Here, we investigate mathematically how the rate of adaptive evolution is affected by the rate of mutation rate phenotype switching. We model an asexual population with two mutation rate phenotypes, non-mutator and mutator. An offspring may switch from its parental phenotype to the other phenotype. We find that switching rates that correspond to so-far empirically described non-genetic systems of inheritance of the mutation rate lead to higher rates of adaptation on both artificial and natural fitness landscapes. These switching rates can maintain within the same individuals both a mutator phenotype and intermediary mutations, a combination that facilitates adaptation. Moreover, non-genetic inheritance increases the proportion of mutators in the population, which in turn increases the probability of hitchhiking of the mutator phenotype with adaptive mutations. This in turns facilitates the acquisition of additional adaptive mutations. Our results rationalize recently observed noise in the expression of proteins that affect the mutation rate and suggest that non-genetic inheritance of this phenotype may facilitate evolutionary adaptive processes.

Mutation rate, selection, and epistasis inferred from RNA virus haplotypes via neural posterior estimation.

RNA viruses are particularly notorious for their high levels of genetic diversity, which is generated through the forces of mutation and natural selection. However, disentangling these two forces is a considerable challenge, and this may lead to widely divergent estimates of viral mutation rates, as well as difficulties in inferring the fitness effects of mutations. Here, we develop, test, and apply an approach aimed at inferring the mutation rate and key parameters that govern natural selection, from haplotype sequences covering full-length genomes of an evolving virus population. Our approach employs neural posterior estimation, a computational technique that applies simulation-based inference with neural networks to jointly infer multiple model parameters. We first tested our approach on synthetic data simulated using different mutation rates and selection parameters while accounting for sequencing errors. Reassuringly, the inferred parameter estimates were accurate and unbiased. We then applied our approach to haplotype sequencing data from a serial passaging experiment with the MS2 bacteriophage, a virus that parasites Escherichia coli. We estimated that the mutation rate of this phage is around 0.2 mutations per genome per replication cycle (95% highest density interval: 0.051-0.56). We validated this finding with two different approaches based on single-locus models that gave similar estimates but with much broader posterior distributions. Furthermore, we found evidence for reciprocal sign epistasis between four strongly beneficial mutations that all reside in an RNA stem loop that controls the expression of the viral lysis protein, responsible for lysing host cells and viral egress. We surmise that there is a fine balance between over- and underexpression of lysis that leads to this pattern of epistasis. To recap, we have developed an approach for joint inference of the mutation rate and selection parameters from full haplotype data with sequencing errors and used it to reveal features governing MS2 evolution.

Conditions that favour cumulative cultural evolution.

The emergence of human societies with complex language and cumulative culture is considered a major evolutionary transition. Why such a high degree of cumulative culture is unique to humans is perplexing given the potential fitness advantages of cultural accumulation. Here, Boyd & Richerson's (1996 Why culture is common, but cultural evolution is rare. Proc. Br. Acad. 88, 77-93) discrete-cultural-trait model is extended to incorporate arbitrarily strong selection; conformist, anti-conformist and unbiased frequency-dependent transmission; random and periodic environmental variation; finite population size; and multiple 'skill levels.' From their infinite-population-size model with success bias and a single skill level, Boyd and Richerson concluded that social learning is favoured over individual learning under a wider range of conditions when social learning is initially common than initially rare. We find that this holds only if the number n of individuals observed by a social learner is sufficiently small, but with a finite population and/or a combination of success-biased and conformist or unbiased transmission, this result holds with larger n. Assuming social learning has reached fixation, the increase in a population's mean skill level is lower if cumulative culture is initially absent than initially present, if population size is finite, or if cultural transmission has a frequency-dependent component. Hence, multiple barriers to cultural accumulation may explain its rarity. This article is part of the theme issue 'Human socio-cultural evolution in light of evolutionary transitions'.

Differential application of cultural practices at the family and individual levels may alter heritability estimates.

Uchiyama et al. emphasize that culture evolves directionally and differentially as a function of selective pressures in different populations. Extending these principles to the level of families, lineages, and individuals exposes additional challenges to estimating heritability. Cultural traits expressed differentially as a function of the genetics whose influence they mask or unmask render inseparable the influences of culture and genetics.

Modeling the evolution of SARS-CoV-2 under non-pharmaceutical interventions and testing.

Background and objectives: Social and behavioral non-pharmaceutical interventions (NPIs), such as mask-wearing, social distancing and travel restrictions, as well as diagnostic tests, have been broadly implemented in response to the COVID-19 pandemic. Epidemiological models and data analysis affirm that wide adoption of NPIs helps to control the pandemic. However, SARS-CoV-2 has extensively demonstrated its ability to evolve. Therefore, it is crucial to examine how NPIs may affect the evolution of the virus. Such evolution could have important effects on the spread and impact of the pandemic. Methodology: We used evo-epidemiological models to examine the effect of NPIs and testing on two evolutionary trajectories for SARS-CoV-2: attenuation and test evasion. Results: Our results show that when stronger measures are taken, selection may act to reduce disease severity. Additionally, the timely application of NPIs could significantly affect the competition between viral strains, favoring the milder strain. Furthermore, a higher testing rate can select for a test-evasive viral strain, even if that strain is less infectious than the detectable competing strain. Importantly, if a less detectable strain evolves, epidemiological metrics such as confirmed daily cases may distort our assessment of the pandemic. Conclusions and implications: Our results highlight the important implications NPIs can have on the evolution of SARS-CoV-2. Lay Summary: We used evo-epidemiological models to examine the effect of non-pharmaceutical interventions and testing on two evolutionary trajectories for SARS-CoV-2: attenuation and test evasion. Our results show that when stronger measures are taken, selection may act to reduce disease severity.

Neural networks enable efficient and accurate simulation-based inference of evolutionary parameters from adaptation dynamics.

The rate of adaptive evolution depends on the rate at which beneficial mutations are introduced into a population and the fitness effects of those mutations. The rate of beneficial mutations and their expected fitness effects is often difficult to empirically quantify. As these 2 parameters determine the pace of evolutionary change in a population, the dynamics of adaptive evolution may enable inference of their values. Copy number variants (CNVs) are a pervasive source of heritable variation that can facilitate rapid adaptive evolution. Previously, we developed a locus-specific fluorescent CNV reporter to quantify CNV dynamics in evolving populations maintained in nutrient-limiting conditions using chemostats. Here, we use CNV adaptation dynamics to estimate the rate at which beneficial CNVs are introduced through de novo mutation and their fitness effects using simulation-based likelihood-free inference approaches. We tested the suitability of 2 evolutionary models: a standard Wright-Fisher model and a chemostat model. We evaluated 2 likelihood-free inference algorithms: the well-established Approximate Bayesian Computation with Sequential Monte Carlo (ABC-SMC) algorithm, and the recently developed Neural Posterior Estimation (NPE) algorithm, which applies an artificial neural network to directly estimate the posterior distribution. By systematically evaluating the suitability of different inference methods and models, we show that NPE has several advantages over ABC-SMC and that a Wright-Fisher evolutionary model suffices in most cases. Using our validated inference framework, we estimate the CNV formation rate at the GAP1 locus in the yeast Saccharomyces cerevisiae to be 10-4.7 to 10-4 CNVs per cell division and a fitness coefficient of 0.04 to 0.1 per generation for GAP1 CNVs in glutamine-limited chemostats. We experimentally validated our inference-based estimates using 2 distinct experimental methods-barcode lineage tracking and pairwise fitness assays-which provide independent confirmation of the accuracy of our approach. Our results are consistent with a beneficial CNV supply rate that is 10-fold greater than the estimated rates of beneficial single-nucleotide mutations, explaining the outsized importance of CNVs in rapid adaptive evolution. More generally, our study demonstrates the utility of novel neural network-based likelihood-free inference methods for inferring the rates and effects of evolutionary processes from empirical data with possible applications ranging from tumor to viral evolution.

Inferring the effective start dates of non-pharmaceutical interventions during COVID-19 outbreaks.

BACKGROUND: During Feb-Apr. 2020, many countries implemented non-pharmaceutical interventions (NPIs), such as school closures and lockdowns, to control the COVID-19 pandemic caused by the SARS-CoV-2 virus. Overall, these interventions seem to have reduced the spread of the pandemic. We hypothesized that the official and effective start dates of NPIs can be noticeably different, for example, due to slow adoption by the population, and that these differences can lead to errors in the estimation of the impact of NPIs. METHODS: SEIR models were fitted to case data from 12 regions to infer the effective start dates of interventions and compare these with the official dates. The impact of NPIs was estimated from the inferred model parameters. RESULTS: We infer mostly late effective start dates of interventions. For example, Italy implemented a lockdown on Mar 11, but we infer the effective start date on Mar 17 (+3.05-2.01 days 95% CI). Moreover, we find that the impact of NPIs can be underestimated if it is assumed they start on their official date. CONCLUSIONS: Differences between the official and effective start of NPIs are likely. Neglecting such differences can lead to underestimation of the impact of NPIs, which could cause decision-makers to escalate interventions and guidelines.

Conformity and content-biased cultural transmission in the evolution of altruism.

The evolution of altruism has been extensively modeled under the assumption of genetic transmission, whereas the dynamics under cultural transmission are less well understood. Previous research has shown that cultural transmission can facilitate the evolution of altruism by increasing (1) the probability of adopting the altruistic phenotype, and (2) assortment between altruists. We incorporate vertical and oblique transmission, which can be conformist or anti-conformist, into models of parental care, sibling altruism, and altruism between individuals that meet assortatively. If oblique transmission is conformist, it becomes easier for altruism to invade a population of non-altruists as the probability of vertical transmission increases. If oblique transmission is anti-conformist, decreasing vertical transmission facilitates invasion by altruism in the assortative meeting model, whereas in other models, there is a trade-off: greater vertical transmission produces greater assortment among genetically related altruists, but lowers the probability of adopting altruism via anti-conformity. Compared to conditions for invasion under genetic transmission, e.g., Hamilton's rule, we show that invasion can be easier with sufficiently strong anti-conformity, and in some models, with sufficiently high assortment even if oblique transmission is conformist. We also explore invasion by an allele A that increases individuals' content bias for altruism, in the absence of other forms of cultural transmission. If costs and benefits combine additively, A invades under previously known conditions. If costs and benefits combine multiplicatively, invasion by A and by altruism become more difficult than in the corresponding additive models.

Annual climatic fluctuations and short-term genetic variation in the eastern spadefoot toad.

In addition to variations on the spatial scale, short- and long-term temporal variations, too, can impose intense selection on the overall genetic diversity and composition of a population. We hypothesized that the allelic composition in populations of the eastern spadefoot toad (Pelobates syriacus) would change among successive years in accordance with the short-term changes in environmental conditions. Surprisingly, the effect of short-term climate fluctuations on genetic composition have rarely been addressed in the literature, and to our knowledge the effect of annual climatic fluctuations have not been considered meaningful. Our findings show that climatic variation among successive years, primarily the amount of rainfall and rainy days, can significantly alter both microsatellite allelic composition and diversity. We suggest that environmental (i.e. fluctuating) selection is differential across the globe, and that its intensity is expected to be greatest in regions where short-term climatic conditions are least stable.

Adaptive Resistance Mutations at Suprainhibitory Concentrations Independent of SOS Mutagenesis.

Emergence of resistant bacteria during antimicrobial treatment is one of the most critical and universal health threats. It is known that several stress-induced mutagenesis and heteroresistance mechanisms can enhance microbial adaptation to antibiotics. Here, we demonstrate that the pathogen Bartonella can undergo stress-induced mutagenesis despite the fact it lacks error-prone polymerases, the rpoS gene and functional UV-induced mutagenesis. We demonstrate that Bartonella acquire de novo single mutations during rifampicin exposure at suprainhibitory concentrations at a much higher rate than expected from spontaneous fluctuations. This is while exhibiting a minimal heteroresistance capacity. The emerged resistant mutants acquired a single rpoB mutation, whereas no other mutations were found in their whole genome. Interestingly, the emergence of resistance in Bartonella occurred only during gradual exposure to the antibiotic, indicating that Bartonella sense and react to the changing environment. Using a mathematical model, we demonstrated that, to reproduce the experimental results, mutation rates should be transiently increased over 1,000-folds, and a larger population size or greater heteroresistance capacity is required. RNA expression analysis suggests that the increased mutation rate is due to downregulation of key DNA repair genes (mutS, mutY, and recA), associated with DNA breaks caused by massive prophage inductions. These results provide new evidence of the hazard of antibiotic overuse in medicine and agriculture.

Non-vertical cultural transmission, assortment and the evolution of cooperation.

Cultural evolution of cooperation under vertical and non-vertical cultural transmission is studied, and conditions are found for fixation and coexistence of cooperation and defection. The evolution of cooperation is facilitated by its horizontal transmission and by an association between social interactions and horizontal transmission. The effect of oblique transmission depends on the horizontal transmission bias. Stable polymorphism of cooperation and defection can occur, and when it does, reduced association between social interactions and horizontal transmission evolves, which leads to a decreased frequency of cooperation and lower population mean fitness. The deterministic conditions are compared to outcomes of stochastic simulations of structured populations. Parallels are drawn with Hamilton's rule incorporating relatedness and assortment.

Adaptive social contact rates induce complex dynamics during epidemics.

Epidemics may pose a significant dilemma for governments and individuals. The personal or public health consequences of inaction may be catastrophic; but the economic consequences of drastic response may likewise be catastrophic. In the face of these trade-offs, governments and individuals must therefore strike a balance between the economic and personal health costs of reducing social contacts and the public health costs of neglecting to do so. As risk of infection increases, potentially infectious contact between people is deliberately reduced either individually or by decree. This must be balanced against the social and economic costs of having fewer people in contact, and therefore active in the labor force or enrolled in school. Although the importance of adaptive social contact on epidemic outcomes has become increasingly recognized, the most important properties of coupled human-natural epidemic systems are still not well understood. We develop a theoretical model for adaptive, optimal control of the effective social contact rate using traditional epidemic modeling tools and a utility function with delayed information. This utility function trades off the population-wide contact rate with the expected cost and risk of increasing infections. Our analytical and computational analysis of this simple discrete-time deterministic strategic model reveals the existence of an endemic equilibrium, oscillatory dynamics around this equilibrium under some parametric conditions, and complex dynamic regimes that shift under small parameter perturbations. These results support the supposition that infectious disease dynamics under adaptive behavior change may have an indifference point, may produce oscillatory dynamics without other forcing, and constitute complex adaptive systems with associated dynamics. Implications for any epidemic in which adaptive behavior influences infectious disease dynamics include an expectation of fluctuations, for a considerable time, around a quasi-equilibrium that balances public health and economic priorities, that shows multiple peaks and surges in some scenarios, and that implies a high degree of uncertainty in mathematical projections.

Cultural evolution of conformity and anticonformity.

Conformist bias occurs when the probability of adopting a more common cultural variant in a population exceeds its frequency, and anticonformist bias occurs when the reverse is true. Conformist and anticonformist bias have been widely documented in humans, and conformist bias has also been observed in many nonhuman animals. Boyd and Richerson used models of conformist and anticonformist bias to explain the evolution of large-scale cooperation, and subsequent research has extended these models. We revisit Boyd and Richerson's original analysis and show that, with conformity based on more than three role models, the evolutionary dynamics can be more complex than previously assumed. For example, we show the presence of stable cycles and chaos under strong anticonformity and the presence of new equilibria when both conformity and anticonformity act at different variant frequencies, with and without selection. We also investigate the case of population subdivision with migration and find that the common claim that conformity can maintain between-group differences is not always true. Therefore, the effect of conformity on the evolution of cooperation by group selection may be more complicated than previously stated. Finally, using Feldman and Liberman's modifier approach, we investigate the conditions under which a rare modifier of the extent of conformity or the number of role models can invade a population. Understanding the dynamics of conformist- and anticonformist-biased transmission may have implications for research on human and nonhuman animal behavior, the evolution of cooperation, and frequency-dependent transmission in general.

A new perspective for mitigation of SARS-CoV-2 infection: priming the innate immune system for viral attack.

The course of infection by SARS-CoV-2 frequently includes a long asymptomatic period, followed in some individuals by an immune dysregulation period that may lead to complications and immunopathology-induced death. This course of disease suggests that the virus often evades detection by the innate immune system. We suggest a novel therapeutic approach to mitigate the infection's severity, probability of complications and duration. We propose that priming an individual's innate immune system for viral attack shortly before it is expected to occur may allow pre-activation of the preferable trajectory of immune response, leading to early detection of the virus. Priming can be carried out, for example, by administering a standard vaccine or another reagent that elicits a broad anti-viral innate immune response. By the time that the expected SARS-CoV-2 infection occurs, activation cascades will have been put in motion and levels of immune factors needed to combat the infection will have been elevated. The infection would thus be cleared faster and with less complication than otherwise, alleviating adverse clinical outcomes at the individual level. Moreover, priming may also mitigate population-level risk by reducing need for hospitalizations and decreasing the infectious period of individuals, thus slowing the spread and reducing the impact of the epidemic. In view of the latter consideration, our proposal may have a significant epidemiological impact even if applied primarily to low-risk individuals, such as young adults, who often show mild symptoms or none, by shortening the period during which they unknowingly infect others. The proposed view is, at this time, an unproven hypothesis. Although supported by robust bio-medical reasoning and multiple lines of evidence, carefully designed clinical trials are necessary.

The evolution of frequency-dependent cultural transmission.

In a model of vertical and oblique cultural transmission of a dichotomous trait, the rates of transmission of each form of the trait are functions of the trait frequency in the population. Sufficient conditions on these functions are derived for a stable trait polymorphism to exist. If the vertical transmission rates are monotone decreasing functions of the trait frequency, a complete global stability analysis is presented. It is also shown that a unique protected polymorphism can be globally stable even though the sufficient conditions are not met. The evolution of frequency-dependent transmission is modeled using modifier theory, and exact conditions are derived for a transmission modifier to invade a population at a stable polymorphism. Finally, the interaction between frequency-dependent selection and frequency-dependent transmission is explored.