Learning the rules of collective cell migration using deep attention networks

Collective, coordinated cellular motions underpin key processes in all multicellular organisms, yet it has been difficult to simultaneously express the ‘rules’ behind these motions in clear, interpretable forms that effectively capture high-dimensional cell-cell interaction dynamics in a manner that is intuitive to the researcher. Here we apply deep attention networks to analyze several canonical living tissues systems and present the underlying collective migration rules for each tissue type using only cell migration trajectory data. We use these networks to learn the behaviors of key tissue types with distinct collective behaviors—epithelial, endothelial, and metastatic breast cancer cells—and show how the results complement traditional biophysical approaches. In particular, we present attention maps indicating the relative influence of neighboring cells to the learned turning decisions of a ‘focal cell’–the primary cell of interest in a collective setting. Colloquially, we refer to this learned relative influence as ‘attention’, as it serves as a proxy for the physical parameters modifying the focal cell’s future motion as a function of each neighbor cell. These attention networks reveal distinct patterns of influence and attention unique to each model tissue. Endothelial cells exhibit tightly focused attention on their immediate forward-most neighbors, while cells in more expansile epithelial tissues are more broadly influenced by neighbors in a relatively large forward sector. Attention maps of ensembles of more mesenchymal, metastatic cells reveal completely symmetric attention patterns, indicating the lack of any particular coordination or direction of interest. Moreover, we show how attention networks are capable of detecting and learning how these rules change based on biophysical context, such as location within the tissue and cellular crowding. That these results require only cellular trajectories and no modeling assumptions highlights the potential of attention networks for providing further biological insights into complex cellular systems.     

Neurocomputational mechanism of controllability inference under a multi-agent setting

Controllability perception significantly influences motivated behavior and emotion and requires an estimation of one’s influence on an environment. Previous studies have shown that an agent can infer controllability by observing contingency between one’s own action and outcome if there are no other outcome-relevant agents in an environment. However, if there are multiple agents who can influence the outcome, estimation of one’s genuine controllability requires exclusion of other agents’ possible influence. Here, we first investigated a computational and neural mechanism of controllability inference in a multi-agent setting. Our novel multi-agent Bayesian controllability inference model showed that other people’s action-outcome contingency information is integrated with one’s own action-outcome contingency to infer controllability, which can be explained as a Bayesian inference. Model-based functional MRI analyses showed that multi-agent Bayesian controllability inference recruits the temporoparietal junction (TPJ) and striatum. Then, this inferred controllability information was leveraged to increase motivated behavior in the vmPFC. These results generalize the previously known role of the striatum and vmPFC in single-agent controllability to multi-agent controllability, and this generalized role requires the TPJ in addition to the striatum of single-agent controllability to integrate both self- and other-related information. Finally, we identified an innate positive bias toward the self during the multi-agent controllability inference, which facilitated behavioral adaptation under volatile controllability. Furthermore, low positive bias and high negative bias were associated with increased daily feelings of guilt. Our results provide a mechanism of how our sense of controllability fluctuates due to other people in our lives, which might be related to social learned helplessness and depression.     

Traits across trophic levels interact to influence parasitoid establishment in biological control releases

A central goal in ecology is to predict what governs a species’ ability to establish in a new environment. One mechanism driving establishment success is individual species’ traits, but the role of trait combinations among interacting species across different trophic levels is less clear. Deliberate or accidental species additions to existing communities provide opportunities to study larger scale patterns of establishment success. Biological control introductions are especially valuable because they contain data on both the successfully established and unestablished species. Here, we used a recent dataset of importation biological control introductions to explore how life‐history traits of 132 parasitoid species and their herbivorous hosts interact to affect parasitoid establishment. We find that of five parasitoid and herbivore traits investigated, one parasitoid trait—host range—weakly predicts parasitoid establishment; parasitoids with higher levels of phylogenetic specialization have higher establishment success, though the effect is marginal. In addition, parasitoids are more likely to establish when their herbivore host has had a shorter residence time. Interestingly, we do not corroborate earlier findings that gregarious parasitoids and endoparasitoids are more likely to establish. Most importantly, we find that life‐history traits of the parasitoid species and their hosts can interact to influence establishment. Specifically, parasitoids with broader host ranges are more likely to establish when the herbivore they have been released to control is also more of a generalist. These results provide insight into how multiple species’ traits and their interactions, both within and across trophic levels, can influence establishment of species of higher trophic levels.     

Subjective optimality in finite sequential decision-making

Many decisions in life are sequential and constrained by a time window. Although mathematically derived optimal solutions exist, it has been reported that humans often deviate from making optimal choices. Here, we used a secretary problem, a classic example of finite sequential decision-making, and investigated the mechanisms underlying individuals’ suboptimal choices. Across three independent experiments, we found that a dynamic programming model comprising subjective value function explains individuals’ deviations from optimality and predicts the choice behaviors under fewer and more opportunities. We further identified that pupil dilation reflected the levels of decision difficulty and subsequent choices to accept or reject the stimulus at each opportunity. The value sensitivity, a model-based estimate that characterizes each individual’s subjective valuation, correlated with the extent to which individuals’ physiological responses tracked stimuli information. Our results provide model-based and physiological evidence for subjective valuation in finite sequential decision-making, rediscovering human suboptimality in subjectively optimal decision-making processes.     

Observing others give & take: A computational account of bystanders’ feelings and actions

Social interactions influence people’s feelings and behavior. Here, we propose that a person’s well-being is influenced not only by interactions they experience themselves, but also by those they observe. In particular, we test and quantify the influence of observed selfishness and observed inequality on a bystanders’ feelings and non-costly punishment decisions. We developed computational models that relate others’ (un)selfish acts to observers’ emotional reactions and punishment decisions. These characterize the rules by which others’ interactions are transformed into bystanders’ reactions, and successfully predict those reactions in out-of-sample participants. The models highlight the impact of two social values—‘selfishness aversion’ and ‘inequality aversion’. As for the latter we find that even small violations from perfect equality have a disproportionately large impact on feelings and punishment. In this age of internet and social media we constantly observe others’ online interactions, in addition to in-person interactions. Quantifying the consequences of such observations is important for predicting their impact on society.     

Status competition,inequality, and fertility: implications for the demographic transition

The role that social status plays in small-scale societies suggests that status may be important for understanding the evolution of human fertility decisions, and for understanding how such decisions play out in modern contexts. This paper explores whether modelling competition for status—in the sense of relative rank within a society—can help shed light on fertility decline and the demographic transition. We develop a model of how levels of inequality and status competition affect optimal investment by parents in the embodied capital (health, strength, and skills) and social status of offspring, focusing on feedbacks between individual decisions and socio-ecological conditions. We find that conditions similar to those in demographic transition societies yield increased investment in both embodied capital and social status, generating substantial decreases in fertility, particularly under conditions of high inequality and intense status competition. We suggest that a complete explanation for both fertility variation in small-scale societies and modern fertility decline will take into account the effects of status competition and inequality.     

Phenotypic plasticity in juvenile jellyfish medusae facilitates effective animal–fluid interaction

Locomotion and feeding in marine animals are intimately linked to the flow dynamics created by specialized body parts. This interaction is of particular importance during ontogeny, when changes in behaviour and scale challenge the organism with shifts in fluid regimes and altered functionality. Previous studies have indicated that Scyphozoan jellyfish ontogeny accommodates the changes in fluid dynamics associated with increasing body dimensions and velocities during development. However, in addition to scale and behaviour that—to a certain degree—underlie the control of the animal, flow dynamics are also dependent on external factors such as temperature. Here, we show phenotypic plasticity in juvenile Aurelia aurita medusae, where morphogenesis is adapted to altered fluid regimes imposed by changes in ambient temperature. In particular, differential proportional growth was found to compensate for temperature-dependent changes in viscous effects, enabling the animal to use adhering water boundary layers as ‘paddles’—and thus economize tissue—at low temperatures, while switching to tissue-dominated propulsion at higher temperatures where the boundary layer thickness is insufficient to serve for paddling. This effect was predicted by a model of animal–fluid interaction and confirmed empirically by flow-field visualization and assays of propulsion efficiency.     

An essential role of the autophagy activating kinase ULK1 in snRNP biogenesis

The biogenesis of small uridine-rich nuclear ribonucleoproteins (UsnRNPs) depends on the methylation of Sm proteins catalyzed by the methylosome and the subsequent action of the SMN complex, which assembles the heptameric Sm protein ring onto small nuclear RNAs (snRNAs). In this sophisticated process, the methylosome subunit pICln (chloride conductance regulatory protein) is attributed to an exceptional key position as an ‘assembly chaperone’ by building up a stable precursor Sm protein ring structure. Here, we show that—apart from its autophagic role—the Ser/Thr kinase ULK1 (Uncoordinated [unc-51] Like Kinase 1) functions as a novel key regulator in UsnRNP biogenesis by phosphorylation of the C-terminus of pICln. As a consequence, phosphorylated pICln is no longer capable to hold up the precursor Sm ring structure. Consequently, inhibition of ULK1 results in a reduction of efficient UsnRNP core assembly. Thus ULK1, depending on its complex formation, exerts different functions in autophagy or snRNP biosynthesis.     

The actin regulators Enabled and Diaphanous direct distinct protrusive behaviors in different tissues during Drosophila development

Actin-based protrusions are important for signaling and migration during development and homeostasis. Defining how different tissues in vivo craft diverse protrusive behaviors using the same genomic toolkit of actin regulators is a current challenge. The actin elongation factors Diaphanous and Enabled both promote barbed-end actin polymerization and can stimulate filopodia in cultured cells. However, redundancy in mammals and Diaphanous’ role in cytokinesis limited analysis of whether and how they regulate protrusions during development. We used two tissues driving Drosophila dorsal closure—migratory leading-edge (LE) and nonmigratory amnioserosal (AS) cells—as models to define how cells shape distinct protrusions during morphogenesis. We found that nonmigratory AS cells produce filopodia that are morphologically and dynamically distinct from those of LE cells. We hypothesized that differing Enabled and/or Diaphanous activity drives these differences. Combining gain- and loss-of-function with quantitative approaches revealed that Diaphanous and Enabled each regulate filopodial behavior in vivo and defined a quantitative “fingerprint”—the protrusive profile—which our data suggest is characteristic of each actin regulator. Our data suggest that LE protrusiveness is primarily Enabled driven, whereas Diaphanous plays the primary role in the AS, and reveal each has roles in dorsal closure, but its robustness ensures timely completion in their absence.     

Relating Land Use and Human Intra-City Mobility

Understanding human mobility patterns—how people move in their everyday lives—is an interdisciplinary research field. It is a question with roots back to the 19^th century that has been dramatically revitalized with the recent increase in data availability. Models of human mobility often take the population distribution as a starting point. Another, sometimes more accurate, data source is land-use maps. In this paper, we discuss how the intra-city movement patterns, and consequently population distribution, can be predicted from such data sources. As a link between land use and mobility, we show that the purposes of people’s trips are strongly correlated with the land use of the trip’s origin and destination. We calibrate, validate and discuss our model using survey data.     

Maintaining Homeostasis by Decision-Making

Living organisms need to maintain energetic homeostasis. For many species, this implies taking actions with delayed consequences. For example, humans may have to decide between foraging for high-calorie but hard-to-get, and low-calorie but easy-to-get food, under threat of starvation. Homeostatic principles prescribe decisions that maximize the probability of sustaining appropriate energy levels across the entire foraging trajectory. Here, predictions from biological principles contrast with predictions from economic decision-making models based on maximizing the utility of the endpoint outcome of a choice. To empirically arbitrate between the predictions of biological and economic models for individual human decision-making, we devised a virtual foraging task in which players chose repeatedly between two foraging environments, lost energy by the passage of time, and gained energy probabilistically according to the statistics of the environment they chose. Reaching zero energy was framed as starvation. We used the mathematics of random walks to derive endpoint outcome distributions of the choices. This also furnished equivalent lotteries, presented in a purely economic, casino-like frame, in which starvation corresponded to winning nothing. Bayesian model comparison showed that—in both the foraging and the casino frames—participants’ choices depended jointly on the probability of starvation and the expected endpoint value of the outcome, but could not be explained by economic models based on combinations of statistical moments or on rank-dependent utility. This implies that under precisely defined constraints biological principles are better suited to explain human decision-making than economic models based on endpoint utility maximization.     

Cooperation and Shared Beliefs about Trust in the Assurance Game

Determinants of cooperation include ingroup vs. outgroup membership, and individual traits, such as prosociality and trust. We investigated whether these factors can be overridden by beliefs about people’s trust. We manipulated the information players received about each other’s level of general trust, “high” or “low”. These levels were either measured (Experiment 1) or just arbitrarily assigned labels (Experiment 2). Players’ choices whether to cooperate or defect in a stag hunt (or an assurance game)—where it is mutually beneficial to cooperate, but costly if the partner should fail to do so—were strongly predicted by what they were told about the other player’s trust label, as well as by what they were told that the other player was told about their own label. Our findings demonstrate the importance for cooperation in a risky coordination game of both first- and second-order beliefs about how much people trust each other. This supports the idea that institutions can influence cooperation simply by influencing beliefs.     

Embodied economics: how bodily information shapes the social coordination dynamics of decision-making

To date, experiments in economics are restricted to situations in which individuals are not influenced by the physical presence of other people. In such contexts, interactions remain at an abstract level, agents guessing what another person is thinking or is about to decide based on money exchange. Physical presence and bodily signals are therefore left out of the picture. However, in real life, social interactions (involving economic decisions or not) are not solely determined by a person''s inference about someone else''s state-of-mind. In this essay, we argue for embodied economics: an approach to neuroeconomics that takes into account how information provided by the entire body and its coordination dynamics influences the way we make economic decisions. Considering the role of embodiment in economics—movements, posture, sensitivity to mimicry and every kind of information the body conveys—makes sense. This is what we claim in this essay which, to some extent, constitutes a plea to consider bodily interactions between agents in social (neuro)economics.     

Tethering-facilitated DNA ‘opening’ and complementary roles of β-hairpin motifs in the Rad4/XPC DNA damage sensor protein

XPC/Rad4 initiates eukaryotic nucleotide excision repair on structurally diverse helix-destabilizing/distorting DNA lesions by selectively ‘opening’ these sites while rapidly diffusing along undamaged DNA. Previous structural studies showed that Rad4, when tethered to DNA, could also open undamaged DNA, suggesting a ‘kinetic gating’ mechanism whereby lesion discrimination relied on efficient opening versus diffusion. However, solution studies in support of such a mechanism were lacking and how ‘opening’ is brought about remained unclear. Here, we present crystal structures and fluorescence-based conformational analyses on tethered complexes, showing that Rad4 can indeed ‘open’ undamaged DNA in solution and that such ‘opening’ can largely occur without one or the other of the β-hairpin motifs in the BHD2 or BHD3 domains. Notably, the Rad4-bound ‘open’ DNA adopts multiple conformations in solution notwithstanding the DNA’s original structure or the β-hairpins. Molecular dynamics simulations reveal compensatory roles of the β-hairpins, which may render robustness in dealing with and opening diverse lesions. Our study showcases how fluorescence-based studies can be used to obtain information complementary to ensemble structural studies. The tethering-facilitated DNA ‘opening’ of undamaged sites and the dynamic nature of ‘open’ DNA may shed light on how the protein functions within and beyond nucleotide excision repair in cells.     

Disparities between observed and predicted impacts of climate change on winter bird assemblages

Understanding how climate change affects the structure and function of communities is critical for gauging its full impact on biodiversity. To date, community-level changes have been poorly documented, owing, in part, to the paucity of long-term datasets. To circumvent this, the use of ‘space-for-time’ substitution—the forecasting of temporal trends from spatial climatic gradients—has increasingly been adopted, often with little empirical support. Here we examine changes from 1975 to 2001 in three community attributes (species richness, body mass and occupancy) for 404 assemblages of terrestrial winter avifauna in North America containing a total of 227 species. We examine the accuracy of space-for-time substitution and assess causal associations between community attributes and observed changes in annual temperature using a longitudinal study design. Annual temperature and all three community attributes increased over time. The trends for the three community attributes differed significantly from the spatially derived predictions, although richness showed broad congruence. Correlations with trends in temperature were found with richness and body mass. In the face of rapid climate change, applying space-for-time substitution as a predictive tool could be problematic with communities developing patterns not reflected by spatial ecological associations.     

Modeling herbivore functional responses causing boom‐bust dynamics following predator removal

Native biodiversity is threatened by invasive species in many terrestrial and marine systems, and conservation managers have demonstrated successes by responding with eradication or control programs. Although invasive species are often the direct cause of threat to native species, ecosystems can react in unexpected ways to their removal or reduction. Here, we use theoretical models to predict boom‐bust dynamics, where the removal of predatory or competitive pressure from a native herbivore results in oscillatory population dynamics (boom‐bust), which can endanger the native species’ population in the short term. We simulate control activities, applied to multiple theoretical three‐species Lotka‐Volterra ecosystem models consisting of vegetation, a native herbivore, and an invasive predator. Based on these communities, we then develop a predictive tool that—based on relative parameter values—predicts whether control efforts directed at the invasive predator will lead to herbivore release followed by a crash. Further, by investigating the different functional responses, we show that model structure, as well as model parameters, are important determinants of conservation outcomes. Finally, control strategies that can mitigate these negative consequences are identified. Managers working in similar data‐poor ecosystems can use the predictive tool to assess the probability that their system will exhibit boom‐bust dynamics, without knowing exact community parameter values.     

Systematically assessing microbiome–disease associations identifies drivers of inconsistency in metagenomic research

Evaluating the relationship between the human gut microbiome and disease requires computing reliable statistical associations. Here, using millions of different association modeling strategies, we evaluated the consistency—or robustness—of microbiome-based disease indicators for 6 prevalent and well-studied phenotypes (across 15 public cohorts and 2,343 individuals). We were able to discriminate between analytically robust versus nonrobust results. In many cases, different models yielded contradictory associations for the same taxon–disease pairing, some showing positive correlations and others negative. When querying a subset of 581 microbe–disease associations that have been previously reported in the literature, 1 out of 3 taxa demonstrated substantial inconsistency in association sign. Notably, >90% of published findings for type 1 diabetes (T1D) and type 2 diabetes (T2D) were particularly nonrobust in this regard. We additionally quantified how potential confounders—sequencing depth, glucose levels, cholesterol, and body mass index, for example—influenced associations, analyzing how these variables affect the ostensible correlation between Faecalibacterium prausnitzii abundance and a healthy gut. Overall, we propose our approach as a method to maximize confidence when prioritizing findings that emerge from microbiome association studies.

The human microbiome has been associated with many aspects of our health, but how many of these associations are truly reproducible? This study attempts to address this question by systematically testing the robustness of 581 microbial features that have been reported as being disease-associated.     

‘Glocal’ Robustness Analysis and Model Discrimination for Circadian Oscillators

To characterize the behavior and robustness of cellular circuits with many unknown parameters is a major challenge for systems biology. Its difficulty rises exponentially with the number of circuit components. We here propose a novel analysis method to meet this challenge. Our method identifies the region of a high-dimensional parameter space where a circuit displays an experimentally observed behavior. It does so via a Monte Carlo approach guided by principal component analysis, in order to allow efficient sampling of this space. This ‘global’ analysis is then supplemented by a ‘local’ analysis, in which circuit robustness is determined for each of the thousands of parameter sets sampled in the global analysis. We apply this method to two prominent, recent models of the cyanobacterial circadian oscillator, an autocatalytic model, and a model centered on consecutive phosphorylation at two sites of the KaiC protein, a key circadian regulator. For these models, we find that the two-sites architecture is much more robust than the autocatalytic one, both globally and locally, based on five different quantifiers of robustness, including robustness to parameter perturbations and to molecular noise. Our ‘glocal’ combination of global and local analyses can also identify key causes of high or low robustness. In doing so, our approach helps to unravel the architectural origin of robust circuit behavior. Complementarily, identifying fragile aspects of system behavior can aid in designing perturbation experiments that may discriminate between competing mechanisms and different parameter sets.     

Stochastic Mesocortical Dynamics and Robustness of Working Memory during Delay-Period

The role of prefronto-mesoprefrontal system in the dopaminergic modulation of working memory during delayed response tasks is well-known. Recently, a dynamical model of the closed-loop mesocortical circuit has been proposed which employs a deterministic framework to elucidate the system’s behavior in a qualitative manner. Under natural conditions, noise emanating from various sources affects the circuit’s functioning to a great extent. Accordingly in the present study, we reformulate the model into a stochastic framework and investigate its steady state properties in the presence of constant background noise during delay-period. From the steady state distribution, global potential landscape and signal-to-noise ratio are obtained which help in defining robustness of the circuit dynamics. This provides insight into the robustness of working memory during delay-period against its disruption due to background noise. The findings reveal that the global profile of circuit’s robustness is predominantly governed by the level of D1 receptor activity and high D1 receptor stimulation favors the working memory-associated sustained-firing state over the spontaneous-activity state of the system. Moreover, the circuit’s robustness is further fine-tuned by the levels of excitatory and inhibitory activities in a way such that the robustness of sustained-firing state exhibits an inverted-U shaped profile with respect to D1 receptor stimulation. It is predicted that the most robust working memory is formed possibly at a subtle ratio of the excitatory and inhibitory activities achieved at a critical level of D1 receptor stimulation. The study also paves a way to understand various cognitive deficits observed in old-age, acute stress and schizophrenia and suggests possible mechanistic routes to the working memory impairments based on the circuit’s robustness profile.     

What are you doing? How active and observational experience shape infants' action understanding

From early in life, infants watch other people''s actions. How do young infants come to make sense of actions they observe? Here, we review empirical findings on the development of action understanding in infancy. Based on this review, we argue that active action experience is crucial for infants'' developing action understanding. When infants execute actions, they form associations between motor acts and the sensory consequences of these acts. When infants subsequently observe these actions in others, they can use their motor system to predict the outcome of the ongoing actions. Also, infants come to an understanding of others’ actions through the repeated observation of actions and the effects associated with them. In their daily lives, infants have plenty of opportunities to form associations between observed events and learn about statistical regularities of others’ behaviours. We argue that based on these two forms of experience—active action experience and observational experience—infants gradually develop more complex action understanding capabilities.