Last month’s parliamentary elections ended Prime Minister Viktor Orbán’s 16-year rule, giving Hungary something more consequential than a change of government: a chance to show the world how to rebuild science after political control. With a two-thirds parliamentary majority, the new leadership has the mandate and the constitutional power to rebuild Hungary’s scientific enterprise around merit, and resistance against future interference. If Hungary gets this right, it will offer a model that matters far beyond its borders.

From disparities in the number of exhibiting artists to auction opportunities, there is evidence of women's under-representation in visual art. Here we explore the exhibition history and auction sales of 65,768 contemporary artists in 20,389 institutions, revealing gender differences in the artist population, exhibitions and auctions. We distinguish between two criteria for gender equity: gender-neutrality, when artists have gender-independent access to exhibition opportunities, and gender-balanced, that strives for gender parity in representation, finding that 58% of institutions are gender-neutral but only 24% are gender-balanced, and that the fraction of man-overrepresented institutions increases with institutional prestige. We define artist's co-exhibition gender to capture the gender inequality of the institutions that an artist exhibits. Finally, we use logistic regression to predict an artist's access to the auction market, finding that co-exhibition gender has a stronger correlation with success than the artist's gender. These results help unveil and quantify the institutional forces that relate to the persistent gender imbalance in the art world.

The brain's connectome1-3 and the vascular system4 are examples of physical networks whose tangible nature influences their structure, layout and, ultimately, their function. The material resources required to build and maintain these networks have inspired decades of research into wiring economy, offering testable predictions about their expected architecture and organization. Here we empirically explore the local branching geometry of a wide range of physical networks, uncovering systematic violations of the long-standing predictions of wiring minimization. This leads to the hypothesis that predicting the true material cost of physical networks requires us to account for their full three-dimensional geometry, resulting in a largely intractable optimization problem. We discover, however, an exact mapping of surface minimization onto high-dimensional Feynman diagrams in string theory5-7, predicting that, with increasing link thickness, a locally tree-like network undergoes a transition into configurations that can no longer be explained by length minimization. Specifically, surface minimization predicts the emergence of trifurcations and branching angles in excellent agreement with the local tree organization of physical networks across a wide range of application domains. Finally, we predict the existence of stable orthogonal sprouts, which are not only prevalent in real networks but also play a key functional role, improving synapse formation in the brain and nutrient access in plants and fungi.

Despite the thousands of genes implicated in age-related phenotypes, effective interventions for aging remain elusive, due to the multifactorial nature of longevity and the interconnectedness of molecular components involved. Here we introduce a network medicine framework to map 2,358 longevity-associated genes onto the human interactome to identify drug-repurposing candidates capable of modulating specific hallmarks of aging. We find that genes associated with each hallmark form a connected subgraph, or hallmark module, allowing us to measure the network proximity of 6,442 compounds to each hallmark. We then introduce a transcription-based metric, pAGE, which evaluates whether drug-induced expression shifts reinforce or counteract known age-related expression changes within each hallmark module. By integrating network proximity and pAGE, we identify drug-repurposing candidates targeting specific hallmarks and provide a falsifiable framework to leverage genomic discoveries for accelerating drug repurposing in longevity. Our findings are interpretable, revealing molecular mechanisms through which drugs modulate hallmarks.

The metaverse is a virtual space enabling interactions beyond geographical boundaries, promising to transform how people engage with each other both in the digital and the physical worlds. The lack of geographical boundaries and travel costs in the metaverse prompts us to ask if the fundamental laws that govern human mobility in the physical world apply. We collected data on avatar movements from Decentraland, along with their network mobility extracted from NFT purchases on Ethereum and Polygon. We find that despite the absence of mobility costs, an individual’s inclination to visit new locations diminishes over time, limiting movement to a small fraction of the metaverse. We also find a lack of correlation between land prices and visitation, a deviation from the patterns characterizing the physical world. Finally, we identify the scaling laws that characterize meta mobility and show that we need to add preferential selection to the existing models to explain quantitative patterns of metaverse mobility. Our ability to predict the characteristics of the emerging meta mobility network implies that the laws governing human mobility are rooted in fundamental patterns of human dynamics, rather than the nature of space and cost of movement.

Somatic mutations accumulate throughout life and have been hypothesized to drive organismal decline. Yet whether these mutations are distributed randomly or whether cells shield their most critical components has remained unresolved. Here we analyze over a million somatic mutations across thirteen human tis-sues, finding that the aging genome exhibits organized vulnerability, captured by the existence of hypo-mutated genes and longevity-associated pathways that have significantly lower mutation burden. Highly connected network hubs are systematically protected from mutation, while peripheral, condition-specific genes accumulate disproportionate burdens. We show that this organized vulnerability arises from the interplay of two independent mechanisms: transcription-coupled repair, and selective filtering. Finally, we validate our findings under experimental mutagenesis, demonstrating intrinsic mechanisms of protection rather than tissue-specific confounders. These findings reframe the somatic mutation hypothesis: organismal decline may not reflect total mutational burden, but where those mutations fall within the cellular network.

The offering of grocery stores is a strong driver of consumer decisions. While highly processed foods such as packaged products, processed meat and sweetened soft drinks have been increasingly associated with unhealthy diets, information on the degree of processing characterizing an item in a store is not straightforward to obtain, limiting the ability of individuals to make informed choices. GroceryDB, a database with over 50,000 food items sold by Walmart, Target and Whole Foods, shows the degree of processing of food items and potential alternatives in the surrounding food environment. The extensive data gathered on ingredient lists and nutrition facts enables a large-scale analysis of ingredient patterns and degrees of processing, categorized by store, food category and price range. Furthermore, it allows the quantification of the individual contribution of over 1,000 ingredients to ultra-processing. GroceryDB makes this information accessible, guiding consumers toward less processed food choices. Information on the degree of processing of food items is key for better consumer choices. GroceryDB is a dataset with more than 50,000 food items sold at major grocery stores in the United States that uses big data to provide information on processing—categorized by store, food category and price range.

Large language models (LLMs) are reshaping how scientific knowledge is accessed and represented. This study evaluates the extent to which popular and frontier LLMs including GPT-4o, Claude 3.5 Sonnet, and Gemini 1.5 Pro recognize scientists, benchmarking their outputs against OpenAlex and Wikipedia. Using a dataset focusing on 100,000 physicists from OpenAlex to evaluate LLM recognition, we uncover substantial disparities: LLMs exhibit selective and inconsistent recognition patterns. Recognition correlates strongly with scholarly impact such as citations, and remains uneven across gender and geography. Women researchers, and researchers from Africa, Asia, and Latin America are significantly underrecognized. We further examine the role of training data provenance, identifying Wikipedia as a potential sources that contributes to recognition gaps. Our findings highlight how LLMs can reflect, and potentially amplify existing disparities in science, underscoring the need for more transparent and inclusive knowledge systems.

SUMMARY:Network medicine leverages the quantification of information flow within sub-cellular networks to elucidate disease etiology and comorbidity, as well as to predict drug efficacy and identify potential therapeutic targets. However, current Network Medicine toolsets often lack computationally efficient data processing pipelines that support diverse scoring functions, network distance metrics, and null models. These limitations hamper their application in large-scale molecular screening, hypothesis testing, and ensemble modeling. To address these challenges, we introduce NetMedPy, a highly efficient and versatile computational package designed for comprehensive Network Medicine analyses. AVAILABILITY AND IMPLEMENTATION:NetMedPy is an open-source Python package under an MIT license. Source code, documentation, and installation instructions can be downloaded from https://github.com/menicgiulia/NetMedPy and https://pypi.org/project/NetMedPy. The package can run on any standard desktop computer or computing cluster.

Network medicine relies on RNA sequencing to infer gene co-expression networks, which are crucial to identify functional gene clusters and gene regulatory interactions, and offer a deeper understanding of disease phenotypes and drug mechanisms. It remains unknown, however, how many samples do we need to make reliable predictions. Here, we propose a power-law model to predict the relationship between the number of inferred significant interactions and sample size, allowing us to quantitatively link sample size to the accuracy of the inferred networks. We apply our model to investigate the effect of sample size on biomarker discovery and differentiation of protein–protein interactions from non-interacting pairs, ultimately unveiling the critical role of data quality in generating meaningful predictions in network medicine.

Biobanks advance biomedical and clinical research by collecting and offering data and biological samples for numerous studies. However, the impact of these repositories varies greatly due to differences in their purpose, scope, governance, and data collected. Here, we computationally identified 2,663 biobanks and their textual mentions in 228,761 scientific articles, 16,210 grants, 15,469 patents, 1,769 clinical trials, and 9,468 public policy documents, helping characterize the academic communities that utilize and support them. We found a strong concentration of biobank-related research on a few diseases, including obesity, Alzheimer's disease, breast cancer, and diabetes. Moreover, collaboration, rather than citation count, shapes the community's recognition of a biobank. We show that, on average, 41.1% of articles fail to reference any of the biobank's reference papers, but 59.6% include a biobank member as a coauthor. Using a generalized linear model, we identified the key factors that contribute to the impact of a biobank, finding that an impactful biobank tends to be more open to external researchers and that quality data-especially linked medical records-as opposed to large data, correlates with a higher impact in science, innovation, and disease. The collected data and findings are accessible through an open-access web application intended to inform strategies to expand access and maximize the value of these resources.

Human-AI coevolution, defined as a process in which humans and AI algorithms continuously influence each other, increasingly characterises our society, but is understudied in artificial intelligence and complexity science literature. Recommender systems and assistants play a prominent role in human-AI coevolution, as they permeate many facets of daily life and influence human choices through online platforms. The interaction between users and AI results in a potentially endless feedback loop, wherein users' choices generate data to train AI models, which, in turn, shape subsequent user preferences. This human-AI feedback loop has peculiar characteristics compared to traditional human-machine interaction and gives rise to complex and often "unintended" systemic outcomes. This paper introduces human-AI coevolution as the cornerstone for a new field of study at the intersection between AI and complexity science focused on the theoretical, empirical, and mathematical investigation of the human-AI feedback loop. In doing so, we: (i) outline the pros and cons of existing methodologies and highlight shortcomings and potential ways for capturing feedback loop mechanisms; (ii) propose a reflection at the intersection between complexity science, AI and society; (iii) provide real-world examples for different human-AI ecosystems; and (iv) illustrate challenges to the creation of such a field of study, conceptualising them at increasing levels of abstraction, i.e., scientific, legal and socio-political.

The brain has long been conceptualized as a network of neurons connected by synapses. However, attempts to describe the connectome using established network science models have yielded conflicting outcomes, leaving the architecture of neural networks unresolved. Here, by performing a comparative analysis of eight experimentally mapped connectomes, we find that their degree distributions cannot be captured by the well-established random or scale-free models. Instead, the node degrees and strengths are well approximated by lognormal distributions, although these lack a mechanistic explanation in the context of the brain. By acknowledging the physical network nature of the brain, we show that neuron size is governed by a multiplicative process, which allows us to analytically derive the lognormal nature of the neuron length distribution. Our framework not only predicts the degree and strength distributions across each of the eight connectomes, but also yields a series of novel and empirically falsifiable relationships between different neuron characteristics. The resulting multiplicative network represents a novel architecture for network science, whose distinctive quantitative features bridge critical gaps between neural structure and function, with implications for brain dynamics, robustness, and synchronization. ### Competing Interest Statement A.-L.B. is the scientific founder of Scipher Medicine, Inc., which applies network medicine to biomarker development.

Background Self-attention mechanisms and token embeddings behind transformers allow the extraction of complex patterns from large datasets, and enhance the predictive power over traditional machine learning models. Yet, being trained to make predictions about individual cells or genes, it is not clear if transformers can learn the inherent interaction patterns between genes, ultimately responsible for their mechanism of action. We use Geneformer, pretrained on single-cell transcriptomes, to ask if transformers can implicitly capture molecular dependencies, including protein-protein interactions (PPIs), allowing us to explore the use of transformers to improve network medicine tasks such as disease gene identification and drug repurposing. Methods We extracted the cosine similarity of gene embeddings and the attention weights contained in Geneformer, allowing us to test if these weights capture experimentally validated protein interactions. Using dilated cardiomyopathy as a case study, we evaluated the effectiveness of the resulting weighted networks in disease module detection and drug repurposing. Results We found that Geneformer displays awareness of experimentally documented Protein-Protein Interactions, exhibiting higher cosine similarity and attention weights for gene pairs with physical interactions. Weighting PPI networks with the cosine similarity and attention weights improved the detection of disease-associated genes and the accuracy of drug repurposing predictions for dilated cardiomyopathy, surpassing the accuracy of unweighted networks. Finally, we find that combining attention weights and cosine similarities with ranking methods enhances drug candidate prioritization for drug repurposing. Conclusions We find that transformers, by implicitly learning the interactions between genes, offer a promising pathway for advancing medicine and drug discovery when integrated with the graph theoretic algorithms used in network medicine. ### Competing Interest Statement A.L.B. is a scientific founder of Scipher Medicine, Inc., which applies network medicine strategies to personalized drug selection. All other authors have no completing interests. ### Funding Statement A.L.B. is supported by the Veterans Affairs Medical Center of Boston Contract #36C24122N0769 and the European Union's Horizon 2020 research and innovation programme under grant agreement No 810115-DYNASNET. ### Author Declarations I confirm all relevant ethical guidelines have been followed, and any necessary IRB and/or ethics committee approvals have been obtained. Yes The details of the IRB/oversight body that provided approval or exemption for the research described are given below: The study uses only public and openly available datasets that were used to train a transformer model, called Geneformer (https://www.nature.com/articles/s41586-023-06139-9). The datasets used in Geneformer are available at https://huggingface.co/datasets/ctheodoris/Genecorpus-30M I confirm that all necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived, and that any patient/participant/sample identifiers included were not known to anyone (e.g., hospital staff, patients or participants themselves) outside the research group so cannot be used to identify individuals. Yes I understand that all clinical trials and any other prospective interventional studies must be registered with an ICMJE-approved registry, such as ClinicalTrials.gov. I confirm that any such study reported in the manuscript has been registered and the trial registration ID is provided (note: if posting a prospective study registered retrospectively, please provide a statement in the trial ID field explaining why the study was not registered in advance). Yes I have followed all appropriate research reporting guidelines, such as any relevant EQUATOR Network research reporting checklist(s) and other pertinent material, if applicable. Yes All raw datasets used in this manuscript are publicly available and directly cited. The source code is available at \url{https://github.com/Barabasi-Lab/Geneformer-NetworkMedicine/tree/main}, along with instructions for recreating and replicating the analyses in the paper. All data produced in this study such as aggregated weight matrices used to create the figures are publicly available at \url{https://huggingface.co/datasets/jspector792/Geneformer_NetworkMedicine}.

Materials, at their essence, are networks defined by homogeneity: uniform bonds, fixed thicknesses, and discrete length scales. Mechanical metamaterials, while representing structurally more diverse microstructures, remain defined by the homogeneity of their unit cells, pore sizes, or repeating features. In contrast, as network science has revealed, real-world and biological systems – from the Internet to the brain – derive their function from broad, multiscale variability in connectivity and link length. Here, we introduce Scale-Rich (SR) metamaterials, a design framework that embeds network heterogeneity into mechanical metamaterials, achieving order-of-magnitude heterogeneity in ligament lengths, thicknesses, and connectivity. Governed by only two parameters, SR networks span orders of magnitude in structural features, overcoming prior constraints in metamaterial design. Translating these network models into physically realizable materials, we use simulations and experiments to show that SR metamaterials exhibit properties inaccessible to traditional single-scale systems, including highly tunable elastic anisotropy, delocalized nonlinear deformation with high energy absorption, and programmable acoustic wave control. This network-science-based paradigm establishes a minimal yet universal framework for engineering multifunctional materials whose mechanical and acoustic behavior emerge directly from scale diversity itself.

References, the mechanism scientists rely on to signal previous knowledge, lately have turned into widely used and misused measures of scientific impact. Yet, when a discovery becomes common knowledge, citations suffer from obliteration by incorporation. This leads to the concept of hidden citation, representing a clear textual credit to a discovery without a reference to the publication embodying it. Here, we rely on unsupervised interpretable machine learning applied to the full text of each paper to systematically identify hidden citations. We find that for influential discoveries hidden citations outnumber citation counts, emerging regardless of publishing venue and discipline. We show that the prevalence of hidden citations is not driven by citation counts, but rather by the degree of the discourse on the topic within the text of the manuscripts, indicating that the more discussed is a discovery, the less visible it is to standard bibliometric analysis. Hidden citations indicate that bibliometric measures offer a limited perspective on quantifying the true impact of a discovery, raising the need to extract knowledge from the full text of the scientific corpus.