That blind spot matters more than many researchers realized. A cell’s fate is not determined solely by the genes it expresses; it is shaped by its position within a tissue, by the density of surrounding cells, by physical constraints, and by local signaling interactions. Conventional RNA velocity methods treat cells as independent entities floating in an abstract expression space, and the result is often inferred dynamics that contradict the actual architecture of the organ being studied. Trajectories may point across anatomical boundaries where no plausible transition exists, or flows may run in directions that make no sense given the known organization of, say, the cerebral cortex. Now a team reporting in Molecular Systems Biology has built a computational framework designed to fix exactly this problem, and in doing so has added something genuinely new: the ability to simulate how perturbing specific genes would redirect cellular dynamics within intact tissue.

The new method, called veloAgent, was developed by Raghavan, Yoon, Fonseca, Li, and Ding and described in a companion News & Views commentary by Jiahao Wu and Suoqin Jin of Wuhan University. At its core, veloAgent takes spliced and unspliced mRNA counts as input and passes them through a variational autoencoder, a type of deep generative neural network that compresses high-dimensional expression data into a compact latent representation. This compression serves two purposes. It reduces the noise and dimensionality that plague raw count data, and it produces a robust summary of each cell’s transcriptional state that downstream models can work with more reliably than with raw counts alone.

From that latent representation, a neural network constructed in conjunction with gene-gene interaction networks predicts kinetic parameters for every cell and every gene: the transcription rate alpha, the splicing rate beta, and the degradation rate gamma. These are the same mechanistic quantities that underlie classical RNA velocity theory, but here they are estimated with the benefit of regulatory structure encoded in known interactions between genes. The initial velocity computed from these parameters is then refined in two stages. The first stage applies velocity smoothing, aligning the directions of transcriptionally similar cells to enforce local coherence and suppress erratic, single-cell noise. The second stage is where veloAgent departs most sharply from its predecessors.

That second stage is an agent-based model, a simulation technique borrowed from fields like epidemiology and economics in which individual entities, or agents, interact according to local rules and collective behavior emerges from those interactions. In veloAgent, each cell is an agent embedded in its spatial microenvironment. The velocity assigned to each cell is adjusted through neighborhood-based weighting that incorporates spatial distance, local cell density, and expression similarity. In practical terms, a cell’s inferred trajectory is constrained by what its actual physical neighbors are doing. Rather than treating spatial information as an auxiliary feature bolted onto the analysis, veloAgent uses it as a primary organizing principle, allowing cell-state transitions to emerge from local interactions rather than being imposed from above. The refined velocity field then supports downstream analyses including spatio-temporal dynamics inference, projection of velocity fields onto low-dimensional spaces and spatial tissue maps, and in silico gene perturbation.

The perturbation capability deserves particular attention because it moves the framework from description toward prediction. Instead of simply altering observed expression values, veloAgent modulates the underlying kinetic parameters directly. Increasing or decreasing alpha for a chosen gene simulates transcriptional activation or repression at the level of the dynamical system itself, after which the induced velocity field is recomputed. The functional impact of each simulated perturbation is quantified by measuring how the directionality of cellular velocity vectors changes, particularly with respect to predefined target states or trajectories. Genes are then ranked by their ability to reorient cellular dynamics toward those targets, yielding a prioritized list of candidate regulators of cell-state transitions. Because the perturbations operate within the learned parameter space, large-scale screening can be performed efficiently without retraining the model, while retaining mechanistic interpretability that purely correlative approaches lack.

The authors validated veloAgent on spatial transcriptomics datasets from multiple platforms and across diverse tissue types. In structured tissues such as the brain, the framework produced trajectories that closely aligned with known anatomical organization, avoiding the inconsistent or anatomically implausible directions that earlier methods generated. In silico perturbation analysis identified key regulatory genes whose simulated modulation reversed progression toward terminal fiber-tract cell states, demonstrating that the model can pinpoint levers that genuinely redirect cellular dynamics. In a breast cancer dataset, veloAgent’s inferred transitions were more consistent with expected tumor progression patterns, correcting misdirected flows produced by competing methods, and further perturbation analysis revealed potential therapeutic targets capable of suppressing malignant progression.

The quantitative benchmarks reinforce the qualitative picture. veloAgent outperformed alternative methods in cross-boundary directionality, fate probability, and velocity confidence, the metrics that capture whether inferred trajectories respect tissue boundaries and how reliably they predict cell fates. Notably, the framework also improved temporal velocity estimation for standard single-cell RNA sequencing data even when no spatial information was available, suggesting that the agent-based refinement and gene-gene regulatory constraints confer benefits that extend beyond the spatial setting. For a field in which velocity estimates have often been fragile and method-dependent, that robustness is a meaningful advance.

The broader significance, as Wu and Jin emphasize in their commentary, is that cellular behavior cannot be understood outside its native environment. Transcriptional state, spatial structure, physical constraint, and local interaction all conspire to determine what a cell becomes. By modeling cells as interacting agents within tissue rather than as independent observations, veloAgent offers a more realistic representation of how cellular processes actually unfold, and it provides a mechanistically grounded platform for simulating interventions before any experiment is run in the laboratory.

Challenges remain, and the commentary is candid about them. Spatial transcriptomics data are frequently collected across multiple tissue sections and experimental conditions, creating batch-correction problems and complicating robust cross-sample comparison of cellular dynamics. Integrating additional modalities such as epigenomics and proteomics would give a fuller picture of the regulatory processes driving cell dynamics. Incorporating causal machine learning frameworks could sharpen both interpretability and predictive power, and directly modeling intercellular signaling, rather than approximating neighborhood effects under spatial constraints, would deepen the mechanistic account of how cells influence one another. Still, the trajectory of the field is clear. The long-term ambition, articulated in recent proposals for building virtual cells with artificial intelligence, is to move beyond descriptive analysis toward predictive models capable of simulating cellular behavior in silico, accelerating therapeutic target discovery and advancing precision medicine. veloAgent represents a concrete step along that path: a tool that not only decodes where cells in a tissue are going, but suggests how researchers might steer them somewhere else.

Subject of Research: Spatially resolved RNA velocity modeling and in silico perturbation of cell-state transitions in tissue

Article Title: Decoding and steering spatially resolved cellular dynamics

Article References: Decoding and steering spatially resolved cellular dynamics. (n.d.). https://doi.org/10.1038/s44320-026-00216-7

Image Credits: AI Generated

DOI: 10.1038/s44320-026-00216-7

Keywords: RNA velocity, spatial transcriptomics, single-cell RNA sequencing, variational autoencoder, agent-based model, cell-state transitions, gene perturbation, computational biology, tumor progression, deep learning, gene regulation, Molecular Systems Biology

CCCvelo represents a significant advancement in the field, as it offers a comprehensive framework for reconstructing the dynamics of CSTs driven by CCC. This innovative tool achieves this by simultaneously optimizing a dynamic CCC signaling network and a latent CST clock. The integration of various processes into a unified model marks a considerable progress toward elucidating the complexities of cell behavior within multicellular systems.

At the core of CCCvelo is a multiscale nonlinear kinetic model that encapsulates the intricacies of intercellular ligand–receptor signaling gradients. This model also accounts for the cascading activation of intracellular transcription factors, ultimately revealing the underlying gene expression dynamics responsible for encoding CSTs. By combining both extrinsic signaling and intrinsic gene regulation, CCCvelo paints a holistic picture of how cellular communication influences developmental trajectories and cellular identities.

To further enhance the model’s capabilities, the researchers developed a unique coevolution learning algorithm dubbed PINN-CELL. This algorithm employs a physics-informed neural network to optimize both model parameters and pseudotemporal ordering concurrently. The dual optimization process enables a more accurate reconstruction of the dynamics at play within the cellular environment. As a result, the application of PINN-CELL offers profound insights into how cell state transitions are orchestrated amidst the noise and complexity inherent in biological systems.

The utility of CCCvelo has been tested on high-resolution ST datasets, including those from mouse cortex, embryonic trunk development, and human prostate cancer. These case studies demonstrate CCCvelo’s prowess in recovering known morphogenetic trajectories while also uncovering how dynamic rewiring of CCC signaling plays a pivotal role in driving CST progression. The implications of these findings extend beyond mere academic curiosity, as they could inform therapeutic strategies in regenerative medicine and cancer treatment.

The use of ST in conjunction with CCCvelo opens new avenues for dissecting the temporal and spatial context of cellular interactions. This enables researchers to identify not just the phases of CSTs but also the underlying communication networks that facilitate these transitions. By capturing the temporal dynamics associated with cell states and transitions, CCCvelo provides a roadmap for understanding more complex biological systems and their emergent properties.

Moreover, CCCvelo’s approach allows researchers to discern subtleties in cell behavior that may have previously gone unnoticed. For example, identifying how certain cell types influence each other’s states through direct communication could reveal potential targets for drug intervention. Understanding these nuanced interactions is critical as therapeutic landscapes increasingly rely on targeting specific signaling pathways rather than broad approaches.

The implications of the model extend to several domains, including developmental biology, cancer research, and regenerative medicine. By tracing the lineage of cell states through the lens of intercellular communication, researchers can begin to delineate the pathways that lead to specific cellular outcomes. This knowledge isn’t only fundamental; it can shape future therapeutic strategies aimed at addressing diseases that arise from dysregulated cell communication.

The CCCvelo framework is especially pertinent in the context of dynamic systems that undergo rapid changes, such as developing embryos or tumor formation. In these scenarios, the ability to capture the temporal progression of cell states can help elucidate the pathways that lead to normal development or pathological conditions. As scientific inquiries into these areas deepen, the relevance of CCCvelo will likely grow, making it an indispensable tool for biologists and medical researchers alike.

Furthermore, the versatility of CCCvelo is noteworthy. It is adaptable to various experimental conditions and can be applied to diverse biological systems across species. This universality enhances its utility across laboratories worldwide, fostering collaborative efforts to unlock the complexities of cell communication and fate determination. By bridging gaps between different research areas, CCCvelo embodies a paradigm shift in understanding multicellular systems.

As the implications of this research unfold, one can anticipate shifts in how cellular networks are visualized and modeled. CCCvelo’s integration of spatial and temporal dimensions provides a new lens through which scientists can scrutinize cellular interactions. In doing so, it not only adds depth to our understanding of CSTs but also challenges existing paradigms and paves the way for novel research questions.

Ultimately, the introduction of CCCvelo as a tool for decoding cell state transitions represents a promising frontier in cellular biology. It encourages a more nuanced appreciation of cellular interactions, signaling dynamics, and the role of communication in shaping cellular outcomes. The ongoing exploration of these interactions offers a treasure trove of potential discoveries, leading to advances in therapeutic strategies as we delve deeper into the molecular underpinnings of life itself.

As we stand on the brink of significant advancements facilitated by technologies like CCCvelo, the future of cellular biology looks bright. The merging of computational methods with experimental data will set the stage for breakthroughs that were previously unimaginable. Such innovations not only advance our understanding but also bring us closer to harnessing the full potential of biology for transformative health solutions, truly underscoring the importance of research in dynamic cellular systems.

Strong collaborative efforts from researchers worldwide will be essential in optimizing CCCvelo and similar tools, enriching our collective understanding even further. The intricate dance of cellular communication is one of the last frontiers in biology, and tools like CCCvelo will surely lead the charge into uncharted territory, where mystery and discovery go hand in hand.

In conclusion, CCCvelo stands as a testament to human ingenuity, representing a leap forward in our quest to decipher the complex language of cellular interactions. It brings us one step closer to unraveling the enigma of how cells communicate and decide their fates, illuminating pathways that could revolutionize personalized medicine and therapeutic interventions. The future is indeed bright, with the potential for breakthroughs that can change the landscape of biology and medicine.

Subject of Research: Cell state transitions driven by dynamic cell–cell communication in spatial transcriptomics.

Article Title: Decoding cell state transitions driven by dynamic cell–cell communication in spatial transcriptomics.

Article References:
Yan, L., Zhang, D. & Sun, X. Decoding cell state transitions driven by dynamic cell–cell communication in spatial transcriptomics.
Nat Comput Sci (2026). https://doi.org/10.1038/s43588-025-00934-2

Image Credits: AI Generated

DOI: https://doi.org/10.1038/s43588-025-00934-2

Keywords: Spatial transcriptomics, cell fate determination, cell state transitions, cell–cell communication, kinetic modeling, CCCvelo, PINN-CELL, signaling networks, lineage tracing, developmental biology.

The motivation behind this research arises from the need to decipher the intricacies of cell differentiation, particularly how early human cells give rise to different somatic and germline lineages. The traditional methods employed for studying these processes often fell short in capturing the complexity of differential gene expression and transitional cell states. This study specifically focused on the induction process of human primordial germ cell-like cells (hPGCLCs) from human pluripotent stem cells, which represent a crucial step towards understanding reproductive cell formation.

scEGOT offers an interpretable and efficient computational approach that stands apart from traditional neural network-based methods. By integrating entropic optimal transport models, this framework allows researchers to construct detailed trajectories of cell differentiation, accurately pinpointing transitional states that other methodologies often overlook. This capability of scEGOT ensures that the nuances of developmental pathways are not only captured but can also be communicated effectively to the scientific community, fostering a deeper understanding of cellular dynamics.

A significant challenge in studying cellular transformation lies in identifying intermediate cell states, which hold vital information regarding the temporal progression of cells as they undergo differentiation. Previous strategies struggled with either defining these states with adequate precision or demanded excessive computational power, which is often prohibitive in large-scale studies. The introduction of scEGOT seeks to address these limitations by providing a rigorous mathematical foundation paired with biologically relevant interpretations.

Dr. Toshiaki Yachimura, the lead researcher on this project, emphasizes the transformative potential of scEGOT. His insights reflect a desire to revolutionize the methodology employed in developmental biology research. By introducing clarity to the process of cell differentiation, scEGOT paves the way for extracting critical information regarding gene regulatory networks that govern these transitions. For instance, through their analysis using scEGOT, researchers uncovered key players in the gene regulatory network revolving around the genes TFAP2A and NKX1-2, crucial for hPGCLC specification.

Moreover, the research team identified that genes like MESP1 and GATA6 play pivotal roles in earlier somatic lineage specification. The findings not only elucidate the molecular underpinnings of early human development but also provide a substantial contribution to the toolkit available for regenerative medicine. By understanding these mechanisms, scientists may eventually unlock new avenues for therapeutic interventions and disease treatment methodologies.

Looking ahead, the versatility of scEGOT allows for further enhancements and applications. Researchers plan to extend the analytical capabilities of this framework to include other single-cell data types such as scATAC-seq, responsible for investigating epigenetic modifications that influence gene expression. This advancement aims to provide a more comprehensive overview of the regulatory networks at play during cell differentiation and may enable a more holistic view of the interplay between various biological molecules.

The discussions surrounding scEGOT highlight the significance of integrating advanced mathematical frameworks with biological insights to tackle fundamental questions in science. As researchers increasingly adopt tools like scEGOT, the implications extend far beyond simple academic curiosity—these advancements hold the promise of accelerating significant discoveries in the field of developmental biology and beyond, bringing us one step closer to unraveling the complex mechanisms that govern cellular life.

Through the combined power of mathematics and biological analysis, scEGOT embodies a new direction for computational tools in biology. It not only enhances the specifics of cell differentiation but also establishes a benchmarking standard for future models that aspire towards high interpretability and computational efficiency. Dr. Yachimura’s work exemplifies a forward-thinking approach in the scientific community, encouraging the exploration of novel mathematical applications to longstanding biological questions.

With this innovative framework entering the scientific literature, the hope is that researchers globally will be inspired to harness its capabilities for their unique research inquiries, ushering in an era where computational biology aids substantially in decoding the complexities of human biology. The significance of this research indicates not just an academic achievement but a substantial contribution to potential medical breakthroughs that could redefine our approach to diseases that have long puzzled researchers.

The convergence of diverse fields and the potential integration of technologies represents the future of scientific exploration. The developments unravelled through scEGOT signify crucial progress, ensuring that the depth of our understanding of cell biology and its implications for human health continues to expand. As we utilize these tools, the scientific community stands on the brink of possibly monumental advances in our quest to elucidate the processes that shape life itself.

Subject of Research: Cells
Article Title: scEGOT: A New Framework in Single-Cell Trajectory Inference
News Publication Date: N/A
Web References: N/A
References: N/A
Image Credits: ASHBi/Kyoto University

Keywords: Computational biology, developmental biology, cell differentiation, single-cell analysis, regenerative medicine, gene regulatory networks, hPGCLCs, epigenetics.