Empowering biomedical discovery with AI agents.

We envision "AI scientists" as systems capable of skeptical learning and reasoning that empower biomedical research through collaborative agents that integrate AI models and biomedical tools with experimental platforms. Rather than taking humans out of the discovery process, biomedical AI agents combine human creativity and expertise with AI's ability to analyze large datasets, navigate hypothesis spaces, and execute repetitive tasks. AI agents are poised to be proficient in various tasks, planning discovery workflows and performing self-assessment to identify and mitigate gaps in their knowledge. These agents use large language models and generative models to feature structured memory for continual learning and use machine learning tools to incorporate scientific knowledge, biological principles, and theories. AI agents can impact areas ranging from virtual cell simulation, programmable control of phenotypes, and the design of cellular circuits to developing new therapies.

Mechanopathology of biofilm-like Mycobacterium tuberculosis cords.

Mycobacterium tuberculosis (Mtb) cultured axenically without detergent forms biofilm-like cords, a clinical identifier of virulence. In lung-on-chip (LoC) and mouse models, cords in alveolar cells contribute to suppression of innate immune signaling via nuclear compression. Thereafter, extracellular cords cause contact-dependent phagocyte death but grow intercellularly between epithelial cells. The absence of these mechanopathological mechanisms explains the greater proportion of alveolar lesions with increased immune infiltration and dissemination defects in cording-deficient Mtb infections. Compression of Mtb lipid monolayers induces a phase transition that enables mechanical energy storage. Agent-based simulations demonstrate that the increased energy storage capacity is sufficient for the formation of cords that maintain structural integrity despite mechanical perturbation. Bacteria in cords remain translationally active despite antibiotic exposure and regrow rapidly upon cessation of treatment. This study provides a conceptual framework for the biophysics and function in tuberculosis infection and therapy of cord architectures independent of mechanisms ascribed to single bacteria.

An ocean of diffusible information.

In the ocean, free-living bacteria exist in a dilute world where direct physical interactions between cells are relatively rare. How then do they exchange genetic information via horizontal gene transfer (HGT)? Lücking et al. have explored the world of marine 'protected extracellular DNA' (peDNA), and find that extracellular vesicles (EVs) are likely to play an important role.

Simulating institutional heterogeneity in sustainability science.

Sustainability outcomes are influenced by the laws and configurations of natural and engineered systems as well as activities in socio-economic systems. An important subset of human activity is the creation and implementation of institutions, formal and informal rules shaping a wide range of human behavior. Understanding these rules and codifying them in computational models can provide important missing insights into why systems function the way they do (static) as well as the pace and structure of transitions required to improve sustainability (dynamic). Here, we conduct a comparative synthesis of three modeling approaches- integrated assessment modeling, engineering-economic optimization, and agent-based modeling-with underexplored potential to represent institutions. We first perform modeling experiments on climate mitigation systems that represent specific aspects of heterogeneous institutions, including formal policies and institutional coordination, and informal attitudes and norms. We find measurable but uneven aggregate impacts, while more politically meaningful distributional impacts are large across various actors. Our results show that omitting institutions can influence the costs of climate mitigation and miss opportunities to leverage institutional forces to speed up emissions reduction. These experiments allow us to explore the capacity of each modeling approach to represent insitutions and to lay out a vision for the next frontier of endogenizing institutional change in sustainability science models. To bridge the gap between modeling, theories, and empirical evidence on social institutions, this research agenda calls for joint efforts between sustainability modelers who wish to explore and incorporate institutional detail, and social scientists studying the socio-political and economic foundations for sustainability transitions.

Using exploratory modeling to challenge narratives of risk governance in Mexico City.

Achieving more sustainable adaptation to social-environmental change demands the transformation of the narratives that provide the rationale for risk governance. These narratives often reflect long-standing beliefs about social and political relationships, ascribe actions and responsibilities, and specify solutions to risk. When such solutions are implemented through material investments in landscapes, these narratives become embedded in physical infrastructure with long legacies. Dominant narratives can mask a range of divergent problem framings. By masking alternatives, narratives can contribute to the persistence of unsustainable governance trajectories. Decision-support tools have begun to represent narratives as drivers of system dynamics; making narratives visible can reveal opportunities for more sustainable governance. We present the results of the project "The Dynamics of Multi-Scalar Adaptation in the Megalopolis", a dynamic, exploratory model of socio-hydrological risks in Mexico City that was designed to both endogenize and simultaneously challenge the dominant narratives that characterize water-risk governance in the city. Qualitative data characterize dominant narratives at city and borough scales. An agent-based model, informed by multicriteria decision analysis and coupled with hydrological, urbanization, and climatic model inputs, permitted the development of exploratory governance scenarios designed to challenge dominant narratives. Scenarios revealed how dominant narratives may contribute to the persistence of vulnerability "hotspots" in the city, despite stated goals of equity and vulnerability alleviation. Participatory workshops with representatives of the city government illustrate how making such narratives visible through exploratory modeling can lead to a questioning of prior assumptions and causal relations, recognition of a need for intersectoral collaboration, and insights into potential management strategies.

Can Generative AI improve social science?

Generative AI that can produce realistic text, images, and other human-like outputs is currently transforming many different industries. Yet it is not yet known how such tools might influence social science research. I argue Generative AI has the potential to improve survey research, online experiments, automated content analyses, agent-based models, and other techniques commonly used to study human behavior. In the second section of this article, I discuss the many limitations of Generative. I examine how bias in the data used to train these tools can negatively impact social science research-as well as a range of other challenges related to ethics, replication, environmental impact, and the proliferation of low-quality research. I conclude by arguing that social scientists can address many of these limitations by creating open-source infrastructure for research on human behavior. Such infrastructure is not only necessary to ensure broad access to high-quality research tools, I argue, but also because the progress of AI will require deeper understanding of the social forces that guide human behavior.

Explaining the evolution of gossip.

Gossip, the exchange of personal information about absent third parties, is ubiquitous in human societies. However, the evolution of gossip remains a puzzle. The current article proposes an evolutionary cycle of gossip and uses an agent-based evolutionary game-theoretic model to assess it. We argue that the evolution of gossip is the joint consequence of its reputation dissemination and selfishness deterrence functions. Specifically, the dissemination of information about individuals' reputations leads more individuals to condition their behavior on others' reputations. This induces individuals to behave more cooperatively toward gossipers in order to improve their reputations. As a result, gossiping has an evolutionary advantage that leads to its proliferation. The evolution of gossip further facilitates these two functions of gossip and sustains the evolutionary cycle.

Collective behavior from surprise minimization.

Collective motion is ubiquitous in nature; groups of animals, such as fish, birds, and ungulates appear to move as a whole, exhibiting a rich behavioral repertoire that ranges from directed movement to milling to disordered swarming. Typically, such macroscopic patterns arise from decentralized, local interactions among constituent components (e.g., individual fish in a school). Preeminent models of this process describe individuals as self-propelled particles, subject to self-generated motion and "social forces" such as short-range repulsion and long-range attraction or alignment. However, organisms are not particles; they are probabilistic decision-makers. Here, we introduce an approach to modeling collective behavior based on active inference. This cognitive framework casts behavior as the consequence of a single imperative: to minimize surprise. We demonstrate that many empirically observed collective phenomena, including cohesion, milling, and directed motion, emerge naturally when considering behavior as driven by active Bayesian inference-without explicitly building behavioral rules or goals into individual agents. Furthermore, we show that active inference can recover and generalize the classical notion of social forces as agents attempt to suppress prediction errors that conflict with their expectations. By exploring the parameter space of the belief-based model, we reveal nontrivial relationships between the individual beliefs and group properties like polarization and the tendency to visit different collective states. We also explore how individual beliefs about uncertainty determine collective decision-making accuracy. Finally, we show how agents can update their generative model over time, resulting in groups that are collectively more sensitive to external fluctuations and encode information more robustly.

Building multiscale models with PhysiBoSS, an agent-based modeling tool.

Multiscale models provide a unique tool for analyzing complex processes that study events occurring at different scales across space and time. In the context of biological systems, such models can simulate mechanisms happening at the intracellular level such as signaling, and at the extracellular level where cells communicate and coordinate with other cells. These models aim to understand the impact of genetic or environmental deregulation observed in complex diseases, describe the interplay between a pathological tissue and the immune system, and suggest strategies to revert the diseased phenotypes. The construction of these multiscale models remains a very complex task, including the choice of the components to consider, the level of details of the processes to simulate, or the fitting of the parameters to the data. One additional difficulty is the expert knowledge needed to program these models in languages such as C++ or Python, which may discourage the participation of non-experts. Simplifying this process through structured description formalisms-coupled with a graphical interface-is crucial in making modeling more accessible to the broader scientific community, as well as streamlining the process for advanced users. This article introduces three examples of multiscale models which rely on the framework PhysiBoSS, an add-on of PhysiCell that includes intracellular descriptions as continuous time Boolean models to the agent-based approach. The article demonstrates how to construct these models more easily, relying on PhysiCell Studio, the PhysiCell Graphical User Interface. A step-by-step tutorial is provided as Supplementary Material and all models are provided at https://physiboss.github.io/tutorial/.

Pioneering bioinformatics with agent-based modelling: an innovative protocol to accurately forecast skin or respiratory allergic reactions to chemical sensitizers.

The assessment of the allergenic potential of chemicals, crucial for ensuring public health safety, faces challenges in accuracy and raises ethical concerns due to reliance on animal testing. This paper presents a novel bioinformatic protocol designed to address the critical challenge of predicting immune responses to chemical sensitizers without the use of animal testing. The core innovation lies in the integration of advanced bioinformatics tools, including the Universal Immune System Simulator (UISS), which models detailed immune system dynamics. By leveraging data from structural predictions and docking simulations, our approach provides a more accurate and ethical method for chemical safety evaluations, especially in distinguishing between skin and respiratory sensitizers. Our approach integrates a comprehensive eight-step process, beginning with the meticulous collection of chemical and protein data from databases like PubChem and the Protein Data Bank. Following data acquisition, structural predictions are performed using cutting-edge tools such as AlphaFold to model proteins whose structures have not been previously elucidated. This structural information is then utilized in subsequent docking simulations, leveraging both ligand-protein and protein-protein interactions to predict how chemical compounds may trigger immune responses. The core novelty of our method lies in the application of UISS-an advanced agent-based modelling system that simulates detailed immune system dynamics. By inputting the results from earlier stages, including docking scores and potential epitope identifications, UISS meticulously forecasts the type and severity of immune responses, distinguishing between Th1-mediated skin and Th2-mediated respiratory allergic reactions. This ability to predict distinct immune pathways is a crucial advance over current methods, which often cannot differentiate between the sensitization mechanisms. To validate the accuracy and robustness of our approach, we applied the protocol to well-known sensitizers: 2,4-dinitrochlorobenzene for skin allergies and trimellitic anhydride for respiratory allergies. The results clearly demonstrate the protocol's ability to differentiate between these distinct immune responses, underscoring its potential for replacing traditional animal-based testing methods. The results not only support the potential of our method to replace animal testing in chemical safety assessments but also highlight its role in enhancing the understanding of chemical-induced immune reactions. Through this innovative integration of computational biology and immunological modelling, our protocol offers a transformative approach to toxicological evaluations, increasing the reliability of safety assessments.

Uncertainty in boundedly rational household adaptation to environmental shocks.

Despite the growing calls to integrate realistic human behavior in sustainability science models, the representative rational agent prevails. This is especially problematic for climate change adaptation that relies on actions at various scales: from governments to individuals. Empirical evidence on individual adaptation to climate-induced hazards reveals diverse behavioral and social factors affecting economic considerations. Yet, implications of replacing the rational optimizer by realistic human behavior in nature-society systems models are poorly understood. Using an innovative evolutionary economic agent-based model we explore different framings regarding household adaptation behavior to floods, leveraging on behavioral data from a household survey in Miami, USA. We find that a representative rational agent significantly overestimates household adaptation diffusion and underestimates damages compared to boundedly rational behavior revealed from our survey. This "adaptation deficit" exhibited by a population of empirically informed agents is explained primarily by diverse "soft" adaptation constraints-awareness, social influences-rather than heterogeneity in financial constraints. Besides initial inequality disproportionally impacting low/medium adaptive capacity households post-flood, our findings suggest that even under a nearly complete adaptation diffusion, adaptation benefits are uneven, with late or less-efficient actions locking households to a path of higher damages, further exacerbating inequalities. Our exploratory modeling reveals that behavioral assumptions shape the uncertainty of physical factors, like exposure and objective effectiveness of flood-proofing measures, traditionally considered crucial in risk assessments. This unique combination of methods facilitates the assessment of cumulative and distributional effects of boundedly rational behavior essential for designing tailored climate adaptation policies, and for equitable sustainability transitions in general.

Fast detection of liver fibrosis with collagen-binding single-nanometer iron oxide nanoparticles via T1-weighted MRI.

SNIO-CBP, a single-nanometer iron oxide (SNIO) nanoparticle functionalized with a type I collagen-binding peptide (CBP), was developed as a T1-weighted MRI contrast agent with only endogenous elements for fast and noninvasive detection of liver fibrosis. SNIO-CBP exhibits 6.7-fold higher relaxivity compared to a molecular gadolinium-based collagen-binding contrast agent CM-101 on a per CBP basis at 4.7 T. Unlike most iron oxide nanoparticles, SNIO-CBP exhibits fast elimination from the bloodstream with a 5.7 min half-life, high renal clearance, and low, transient liver enhancement in healthy mice. We show that a dose of SNIO-CBP that is 2.5-fold lower than that for CM-101 has comparable imaging efficacy in rapid (within 15 min following intravenous injection) detection of hepatotoxin-induced liver fibrosis using T1-weighted MRI in a carbon tetrachloride-induced mouse liver injury model. We further demonstrate the applicability of SNIO-CBP in detecting liver fibrosis in choline-deficient L-amino acid-defined high-fat diet mouse model of nonalcoholic steatohepatitis. These results provide a platform with potential for the development of high relaxivity, gadolinium-free molecular MRI probes for characterizing chronic liver disease.

Spreading processes with mutations over multilayer networks.

A key scientific challenge during the outbreak of novel infectious diseases is to predict how the course of the epidemic changes under countermeasures that limit interaction in the population. Most epidemiological models do not consider the role of mutations and heterogeneity in the type of contact events. However, pathogens have the capacity to mutate in response to changing environments, especially caused by the increase in population immunity to existing strains, and the emergence of new pathogen strains poses a continued threat to public health. Further, in the light of differing transmission risks in different congregate settings (e.g., schools and offices), different mitigation strategies may need to be adopted to control the spread of infection. We analyze a multilayer multistrain model by simultaneously accounting for i) pathways for mutations in the pathogen leading to the emergence of new pathogen strains, and ii) differing transmission risks in different settings, modeled as network layers. Assuming complete cross-immunity among strains, namely, recovery from any infection prevents infection with any other (an assumption that will need to be relaxed to deal with COVID-19 or influenza), we derive the key epidemiological parameters for the multilayer multistrain framework. We demonstrate that reductions to existing models that discount heterogeneity in either the strain or the network layers may lead to incorrect predictions. Our results highlight that the impact of imposing/lifting mitigation measures concerning different contact network layers (e.g., school closures or work-from-home policies) should be evaluated in connection with their effect on the likelihood of the emergence of new strains.

Stopping Aspirin Within 1 Month After Stenting for Ticagrelor Monotherapy in Acute Coronary Syndrome: The T-PASS Randomized Noninferiority Trial.

BACKGROUND: Stopping aspirin within 1 month after implantation of a drug-eluting stent for ticagrelor monotherapy has not been exclusively evaluated for patients with acute coronary syndrome. The aim of this study was to investigate whether ticagrelor monotherapy after <1 month of dual antiplatelet therapy (DAPT) is noninferior to 12 months of ticagrelor-based DAPT for adverse cardiovascular and bleeding events in patients with acute coronary syndrome. METHODS: In this randomized, open-label, noninferiority trial, 2850 patients with acute coronary syndrome who underwent drug-eluting stent implantation at 24 centers in South Korea were randomly assigned (1:1) to receive either ticagrelor monotherapy (90 mg twice daily) after <1 month of DAPT (n=1426) or 12 months of ticagrelor-based DAPT (n=1424) between April 24, 2019, and May 31, 2022. The primary end point was the net clinical benefit as a composite of all-cause death, myocardial infarction, definite or probable stent thrombosis, stroke, and major bleeding at 1 year after the index procedure in the intention-to-treat population. Key secondary end points were the individual components of the primary end point. RESULTS: Among 2850 patients who were randomized (mean age, 61 years; 40% ST-segment-elevation myocardial infarction), 2823 (99.0%) completed the trial. Aspirin was discontinued at a median of 16 days (interquartile range, 12-25 days) in the group receiving ticagrelor monotherapy after <1 month of DAPT. The primary end point occurred in 40 patients (2.8%) in the group receiving ticagrelor monotherapy after <1-month DAPT, and in 73 patients (5.2%) in the ticagrelor-based 12-month DAPT group (hazard ratio, 0.54 [95% CI, 0.37-0.80]; P<0.001 for noninferiority; P=0.002 for superiority). This finding was consistent in the per-protocol population as a sensitivity analysis. The occurrence of major bleeding was significantly lower in the ticagrelor monotherapy after <1-month DAPT group compared with the 12-month DAPT group (1.2% versus 3.4%; hazard ratio, 0.35 [95% CI, 0.20-0.61]; P<0.001). CONCLUSIONS: This study provides evidence that stopping aspirin within 1 month for ticagrelor monotherapy is both noninferior and superior to 12-month DAPT for the 1-year composite outcome of death, myocardial infarction, stent thrombosis, stroke, and major bleeding, primarily because of a significant reduction in major bleeding, among patients with acute coronary syndrome receiving drug-eluting stent implantation. Low event rates, which may suggest enrollment of relatively non-high-risk patients, should be considered in interpreting the trial. REGISTRATION: URL: https://www.clinicaltrials.gov; Unique identifier: NCT03797651.

Regulation of Erythropoiesis by the Hypoxia-Inducible Factor Pathway: Effects of Genetic and Pharmacological Perturbations.

Red blood cells transport O2 from the lungs to body tissues. Hypoxia stimulates kidney cells to secrete erythropoietin (EPO), which increases red cell mass. Hypoxia-inducible factors (HIFs) mediate EPO gene transcriptional activation. HIF-α subunits are subject to O2-dependent prolyl hydroxylation and then bound by the von Hippel-Lindau protein (VHL), which triggers their ubiquitination and proteasomal degradation. Mutations in the genes encoding EPO, EPO receptor, HIF-2α, prolyl hydroxylase domain protein 2 (PHD2), or VHL cause familial erythrocytosis. In addition to O2, α-ketoglutarate is a substrate for PHD2, and analogs of α-ketoglutarate inhibit hydroxylase activity. In phase III clinical trials evaluating the treatment of anemia in chronic kidney disease, HIF prolyl hydroxylase inhibitors were as efficacious as darbepoetin alfa in stimulating erythropoiesis. However, safety concerns have arisen that are focused on thromboembolism, which is also a phenotypic manifestation of VHL or HIF-2α mutation, suggesting that these events are on-target effects of HIF prolyl hydroxylase inhibitors.

Forward-reverse mutation cycles in cancer cell lines under chemical treatments.

Spontaneous forward-reverse mutations were reported by us earlier in clinical samples from various types of cancers and in HeLa cells under normal culture conditions. To investigate the effects of chemical stimulations on such mutation cycles, the present study examined single nucleotide variations (SNVs) and copy number variations (CNVs) in HeLa and A549 cells exposed to wogonin-containing or acidic medium. In wogonin, both cell lines showed a mutation cycle during days 16-18. In acidic medium, both cell lines displayed multiple mutation cycles of different magnitudes. Genomic feature colocalization analysis suggests that CNVs tend to occur in expanded and unstable regions, and near promoters, histones, and non-coding transcription sites. Moreover, phenotypic variations in cell morphology occurred during the forward-reverse mutation cycles under both types of chemical treatments. In conclusion, chemical stresses imposed by wogonin or acidity promoted cyclic forward-reverse mutations in both HeLa and A549 cells to different extents.

The Moral Psychology of Artificial Intelligence.

Moral psychology was shaped around three categories of agents and patients: humans, other animals, and supernatural beings. Rapid progress in artificial intelligence has introduced a fourth category for our moral psychology to deal with: intelligent machines. Machines can perform as moral agents, making decisions that affect the outcomes of human patients or solving moral dilemmas without human supervision. Machines can be perceived as moral patients, whose outcomes can be affected by human decisions, with important consequences for human-machine cooperation. Machines can be moral proxies that human agents and patients send as their delegates to moral interactions or use as a disguise in these interactions. Here we review the experimental literature on machines as moral agents, moral patients, and moral proxies, with a focus on recent findings and the open questions that they suggest.

A functional RNA-origami as direct thrombin inhibitor with fast-acting and specific single-molecule reversal agents in vivo model.

Injectable anticoagulants are widely used in medical procedures to prevent unwanted blood clotting. However, many lack safe, effective reversal agents. Here, we present new data on a previously described RNA origami-based, direct thrombin inhibitor (HEX01). We describe a new, fast-acting, specific, single-molecule reversal agent (antidote) and present in vivo data for the first time, including efficacy, reversibility, preliminary safety, and initial biodistribution studies. HEX01 contains multiple thrombin-binding aptamers appended on an RNA origami. It exhibits excellent anticoagulation activity in vitro and in vivo. The new single-molecule, DNA antidote (HEX02) reverses anticoagulation activity of HEX01 in human plasma within 30 s in vitro and functions effectively in a murine liver laceration model. Biodistribution studies of HEX01 in whole mice using ex vivo imaging show accumulation mainly in the liver over 24 h and with 10-fold lower concentrations in the kidneys. Additionally, we show that the HEX01/HEX02 system is non-cytotoxic to epithelial cell lines and non-hemolytic in vitro. Furthermore, we found no serum cytokine response to HEX01/HEX02 in a murine model. HEX01 and HEX02 represent a safe and effective coagulation control system with a fast-acting, specific reversal agent showing promise for potential drug development.

Surface chemistry engineered selenium nanoparticles as bactericidal and immuno-modulating dual-functional agents for combating methicillin-resistant Staphylococcus aureus Infection.

Because of the extremely complexed microenvironment of drug-resistant bacterial infection, nanomaterials with both bactericidal and immuno-modulating activities are undoubtedly the ideal modality for overcoming drug resistance. Herein, we precisely engineered the surface chemistry of selenium nanoparticles (SeNPs) using neutral (polyvinylpyrrolidone-PVP), anionic (letinan-LET) and cationic (chitosan-CS) surfactants. It was found that surface chemistry greatly influenced the bioactivities of functionalized SeNPs, their interactions with methicillin-resistant Staphylococcus aureus (MRSA), immune cells and metabolisms. LET-functionalized SeNPs with distinct metabolisms exhibited the best inhibitory efficacy compared to other kinds of SeNPs against MRSA through inducing robust ROS generation and damaging bacterial cell wall. Meanwhile, only LET-SeNPs could effectively activate natural kill (NK) cells, and enhance the phagocytic capability of macrophages and its killing activity against bacteria. Furthermore, in vivo studies suggested that LET-SeNPs treatment highly effectively combated MRSA infection and promoted wound healing by triggering much more mouse NK cells, CD8+ and CD4+ T lymphocytes infiltrating into the infected area at the early stage to efficiently eliminate MRSA in the mouse model. This study demonstrates that the novel functionalized SeNP with dual functions could serve as an effective antibacterial agent and could guide the development of next generation antibacterial agents.

How digital media drive affective polarization through partisan sorting.

Politics has in recent decades entered an era of intense polarization. Explanations have implicated digital media, with the so-called echo chamber remaining a dominant causal hypothesis despite growing challenge by empirical evidence. This paper suggests that this mounting evidence provides not only reason to reject the echo chamber hypothesis but also the foundation for an alternative causal mechanism. To propose such a mechanism, the paper draws on the literatures on affective polarization, digital media, and opinion dynamics. From the affective polarization literature, we follow the move from seeing polarization as diverging issue positions to rooted in sorting: an alignment of differences which is effectively dividing the electorate into two increasingly homogeneous megaparties. To explain the rise in sorting, the paper draws on opinion dynamics and digital media research to present a model which essentially turns the echo chamber on its head: it is not isolation from opposing views that drives polarization but precisely the fact that digital media bring us to interact outside our local bubble. When individuals interact locally, the outcome is a stable plural patchwork of cross-cutting conflicts. By encouraging nonlocal interaction, digital media drive an alignment of conflicts along partisan lines, thus effacing the counterbalancing effects of local heterogeneity. The result is polarization, even if individual interaction leads to convergence. The model thus suggests that digital media polarize through partisan sorting, creating a maelstrom in which more and more identities, beliefs, and cultural preferences become drawn into an all-encompassing societal division.