Gene regulatory networks embody the intricate relationships among genes that orchestrate cellular behavior, development, and response to environmental stimuli. Despite scRNA-seq technology allowing researchers to capture gene expression profiles at single-cell resolution, extracting meaningful regulatory connections from this high-dimensional, sparse, and noisy data remains a daunting task. Conventional GRN inference methods frequently rely on examining pairwise relationships between genes, an approach that, while informative, fails to fully capture the global structural dependencies and systemic connectivity that characterize biological networks.

ZINB-GRAN pioneers a shift from pairwise focus to holistic network construction by reformulating GRN inference as a link prediction problem within a weighted gene co-expression network (WGCN). Initially, the model constructs a WGCN from the scRNA-seq count matrix, capturing gene expression correlations across individual cells while preserving crucial biological signals. This WGCN serves as a foundational prior graph, reflecting preliminary insights into gene-gene co-expression patterns, which are integral to inferring regulatory interactions.

A core component of ZINB-GRAN is its use of graph convolutional networks (GCNs) embedded within a graph autoencoder (GAE) framework. The GCN encoder assimilates topological features from the WGCN, learning latent, low-dimensional representations of genes that encapsulate their regulatory context within the global network. Subsequently, a decoder function scores these gene embeddings to reconstruct the GRN, estimating the likelihood of regulatory links and unveiling the underlying gene interplay.

Crucially, ZINB-GRAN incorporates a zero-inflated negative binomial (ZINB) distributional prior to tackle the unique statistical properties of scRNA-seq data. Single-cell expression datasets are notoriously sparse, riddled with dropout events where genuine gene expression appears as zero due to technical limitations. By integrating the ZINB prior, the model statistically mirrors the excess zero inflation and overdispersion inherent in single-cell gene counts, ensuring that latent gene representations align with biologically plausible expression distributions.

This prior is cleverly transformed into a continuous latent space through rigorous sampling, normalization, and Gaussian perturbations. An adversarial training protocol is then employed to align the learned latent embeddings with this continuous prior distribution, fostering biologically consistent gene representations. Such adversarial regularization minimizes discrepancies between the model’s inferred gene regulatory structure and the expected statistical behavior of single-cell data, enhancing robustness and interpretability.

The training of ZINB-GRAN optimizes two intertwined objectives: the supervised task of reconstructing the gene regulatory network and the adversarial objective aligning latent representations with the ZINB prior. This joint optimization strategy enables the model to effectively navigate the noise and sparsity characteristic of scRNA-seq datasets, yielding high-fidelity regulatory networks capable of reflecting true biological complexity rather than technical artifacts.

Benchmarking studies on simulated and real-world datasets illustrate the superior performance of ZINB-GRAN over existing GRN inference methods. Not only does it excel in reconstructing accurate regulatory network topologies, but it also demonstrates enhanced capacity to identify biologically meaningful gene interactions, outperforming traditional correlation-based and regression techniques. These advancements mark a significant step forward in the computational reconstruction of cellular regulatory frameworks.

Applications of ZINB-GRAN to human peripheral blood mononuclear cells (PBMCs) highlight its potential to uncover cell type-specific regulatory networks, providing a granular understanding of immune cell regulation. The framework effectively identifies key transcription factors and regulatory modules underpinning immune function, supporting translational research into immune responses and disease mechanisms.

Additionally, the framework’s application to triple-negative breast cancer datasets underscores its ability to dissect complex disease mechanisms at the regulatory level. By pinpointing critical regulatory factors associated with cancer progression and heterogeneity, ZINB-GRAN offers a promising avenue for discovering novel biomarkers and therapeutic targets in oncology.

The novel integration of global network topology learning with biologically informed statistical priors in ZINB-GRAN represents a paradigm shift in single-cell gene regulatory network inference. Its sophisticated graph-based architecture, combined with rigorous statistical modeling and adversarial training, not only enhances the accuracy of inferred networks but also ensures their biological validity and interpretability, addressing a major bottleneck in systems biology.

This methodology opens exciting possibilities for the study of cellular regulatory mechanisms across diverse biological contexts, enabling researchers to more reliably reconstruct the regulatory wiring diagrams that dictate cell fate decisions, development, and disease pathology. ZINB-GRAN stands as a versatile, scalable tool, positioning itself at the forefront of computational biology and precision medicine.

Future directions for ZINB-GRAN may include integration with multi-omics data to further enrich regulatory inference, extension to temporal and spatial single-cell datasets, and real-time application in personalized therapeutic strategies. Its capacity to model complex, high-dimensional biological data in a statistically sound, biologically coherent manner propels it toward becoming an indispensable framework in modern genomics research.

In summary, ZINB-GRAN offers a groundbreaking advancement in the reconstruction of gene regulatory networks from single-cell RNA sequencing data by effectively integrating graph convolutional networks with a zero-inflated negative binomial prior, adversarial training, and a global network perspective, heralding a new era of precision and insight in the decoding of cellular regulation.

Subject of Research: Not applicable

Article Title: ZINB-GRAN: A ZINB-prior graph adversarial framework for gene regulatory network inference from scRNA-seq data

News Publication Date: 11-Jun-2026

Web References: http://dx.doi.org/10.70401/cbm.2026.0017

Image Credits: © Jianping Zhao, Junfeng Xia, Chunhou Zheng, et al. This is an Open Access article licensed under a Creative Commons Attribution 4.0 International License.

Keywords: gene regulatory networks, single-cell RNA sequencing, graph convolutional networks, zero-inflated negative binomial, graph adversarial learning, gene co-expression networks, biological network inference, deep learning, computational biology, data sparsity, transcriptional regulation, cancer genomics

The research hinged upon cutting-edge spatial transcriptomic technology, specifically the Xenium platform by 10X Genomics, which allowed for high-resolution gene expression profiling of thousands of cells within coronal sections of the MeA. The authors prepared tissue slices spanning the range from −0.7 mm to −2.06 mm posterior to the bregma in murine brains, tagging VMH-projecting neurons with a retrograde fluorescent label via AAVretro-hSyn-mCherry injections into the VMH. This enabled precise mapping and phenotyping of the MeA neurons that interface directly with this hypothalamic nucleus, a pivotal hub in energy and endocrine homeostasis.

Clustering analyses of over 21,600 MeA cells revealed 15 distinct cellular populations encompassing both neuronal and non-neuronal types. Within the neurons, unsupervised re-clustering isolated 20 discrete neural clusters, unveiling heterogeneity in neurotransmitter identity and topographic distribution. Notably, a major GABAergic population expressing Vgat (Slc32a1) dominated, identified distinctly through uniform manifold approximation and projection (UMAP), contrasting with three separate groups of glutamatergic neurons each characterized by exclusive expression of Vglut1 (Slc17a7), Vglut2 (Slc17a6), or co-expression of both. These excitatory neurons displayed a striking ventral MeA localization, segregated along the anterior-posterior axis, with Vglut2 concentrated anteriorly and Vglut1 posteriorly, while inhibitory neurons predominantly occupied the dorsal MeA.

Importantly, the distribution of stress-induced immediate early gene expression (FOS) spanned both dorsal and ventral MeA regions, indicating that stress activates a broad spectrum of excitatory and inhibitory MeA neurons. This finding underscored the functional relevance of discrete neurotransmitter populations in the MeA’s overall response to stress stimuli, suggesting multilayered regulatory mechanisms at play in neural circuits controlling systemic glucose dynamics.

Integration of transcriptomic data with retrograde fluorescent labeling pinpointed 305 VMH-projecting MeA neurons, which were predominately glutamatergic, composing approximately 74% of this projection pool. These neurons were mostly concentrated in clusters enriched for Vglut2+ neurons (clusters 3 and 4), as well as those expressing Vglut1 and mixed Vglut1/2. Intriguingly, a subset of GABAergic neurons (cluster 11) also contributed to this projection, highlighting a dual excitatory-inhibitory input from the MeA to the VMH. Differential gene expression analyses revealed that these VMH-projecting neurons express unique gene signatures associated with metabolic and glycaemic regulation, which also intersect with human genetic loci linked to type 2 diabetes and body weight regulation.

To validate the anatomical connectivity suggested by transcriptomics, the team employed Cre-dependent viral tracing using synaptophysin-mCherry in genetically defined mouse lines expressing Cre recombinase in glutamatergic (Vglut2-cre) or GABAergic (Vgat-cre) neurons. Synaptophysin, a presynaptic vesicle protein, allowed visualization of axonal terminals emanating from MeA neurons. Robust labeling was observed in the VMH from both genetically targeted populations, conclusively demonstrating that excitatory and inhibitory MeA neurons send direct projections to this hypothalamic nucleus, establishing a dual-modality output channel to this critical metabolic center.

Functional interrogation of these identified circuits using chemogenetics further solidified their physiological impact. Activation of MeA glutamatergic neurons via CamK2a-driven hM3DGq receptor expression, as well as the stimulation of GABAergic neurons using a Dlx promoter-driven hM3DGq receptor system, both resulted in significant elevations of blood glucose following clozapine-N-oxide (CNO) administration. Control animals expressing mCherry alone did not exhibit these changes, underscoring the causal role of MeA neuronal activity in modulating systemic glucose levels during stress. These findings illuminate a bidirectional control mechanism in which both excitation and inhibition from the MeA shape hypothalamic output to peripheral metabolic organs.

The molecular profile of VMH-projecting MeA neurons also opens avenues for understanding the genetics of metabolic disorders. By integrating Human Genetic Evidence (HuGE) scores associated with glycaemic traits, the authors linked gene expression patterns within these cells to allelic variants influencing glucose homeostasis and diabetes risk. This convergence of transcriptomic and genetic data highlights the MeA^VMH neuronal phenotype as a critical node for potential therapeutic targeting in stress-related metabolic dysfunction.

Collectively, this comprehensive study redefines the medial amygdala beyond its classical role in emotional behavior, positioning it as a nuanced integrator of neural circuits governing glucose regulation during stress. The discovery of mixed glutamatergic and GABAergic projections to the VMH, coupled with their functional validation, underscores an elegant interplay between excitatory and inhibitory signaling in the brain’s control over peripheral metabolism. These insights pave the way for deeper investigations into how emotional and metabolic states coalesce at the level of discrete neural ensembles.

Moreover, this research exemplifies the power of spatial transcriptomics combined with viral circuit tracing and chemogenetics to unravel complex brain-body communication pathways. The ability to profile gene expression in situ with single-cell resolution, alongside pinpointing functional connectivity, allows for unprecedented dissection of neural circuits that coordinate systemic physiological responses. It heralds a new era in neuroscience focused on connecting genomic determinants with behavioral and metabolic phenotypes.

The implications for understanding stress-related disorders, including diabetes and obesity, are profound. Stress is a well-known precipitant of hyperglycaemia and metabolic dysregulation, yet the precise neurobiological substrates have remained elusive. This study’s identification of a specific amygdala-to-hypothalamus projection that modulates blood glucose reveals potential targets for intervention that could decouple harmful metabolic sequelae from stress exposure.

In future directions, deciphering the downstream targets and signaling cascades within the hypothalamus and peripheral organs influenced by these MeA neurons will be critical. Additionally, exploring how chronic stress or pathological conditions alter this circuit’s function and gene expression profiles may shed light on mechanisms driving metabolic disease progression. The integration of multi-omics approaches with in vivo functional studies promises to further expand our understanding of brain-metabolism crosstalk.

In summary, the meticulous characterization of MeA^VMH neurons as heterogeneous populations of glutamatergic and GABAergic cells projecting to the VMH, together with their demonstrable influence on blood glucose during stress, represents a seminal advance in neuroendocrinology. These findings unravel a vital neural pathway that links emotion-processing centers with metabolic control, enhancing our grasp of how the brain orchestrates complex physiological adaptations to stress.

Subject of Research: Neural circuits connecting the medial amygdala to the ventromedial hypothalamus and their role in regulating stress-induced blood glucose responses.

Article Title: Amygdala–liver signalling orchestrates glycaemic responses to stress.

Article References:
Carty, J.R.E., Devarakonda, K., O’Connor, R.M. et al. Amygdala–liver signalling orchestrates glycaemic responses to stress. Nature (2025). https://doi.org/10.1038/s41586-025-09420-1