Traditional methods in single-cell biology have offered increasingly detailed developmental maps by tracing cellular trajectories using RNA velocity approaches. These methods estimate the direction of a cell’s progression based on immature and mature RNA ratios, effectively capturing the velocity vector of cellular states within developmental landscapes. However, until now, the molecular underpinnings that mechanistically dictate these trajectories—particularly the gene regulatory networks (GRNs) that intricately control gene expression—have largely been studied in isolation. RegVelo bridges this critical gap by marrying the dynamics of RNA velocity with the regulatory circuitry governing gene interactions, forming a holistic computational framework.

The core innovation of RegVelo lies in its ability to treat genes not as isolated entities but as components of an interconnected network, where transcription factors and other regulatory genes can exert activating or repressive influences on each other. By embedding these regulatory relationships within a deep learning model, RegVelo simultaneously deciphers how cells transition from one developmental state to another and identifies the molecular drivers that orchestrate those transitions. This dual insight enables predictions about cellular fate that are both trajectory-aware and regulatory-informed—a feat that previous methods, limited to either trajectory or regulation alone, could not accomplish.

At the helm of this cutting-edge work, Prof. Fabian J. Theis, Director of the Computational Health Center at Helmholtz Munich and Professor at the Technical University of Munich, emphasizes the transformative nature of RegVelo. “Our framework does more than chart cellular pathways,” Theis explains. “It illuminates the regulatory relationships that actively shape those paths, providing a dynamic map of gene interactions in action as cells develop. This allows us not only to observe but to interrogate and simulate the genetic ‘engines’ driving development.”

The genesis of RegVelo is itself a testament to scientific synergy, arising from the union of high-resolution experimental data and sophisticated computational modeling. Tatjana Sauka-Spengler, an Investigator at the Stowers Institute who transitioned from the University of Oxford, contributed richly detailed gene regulatory circuits from her laboratory’s pioneering research on cranial neural crest cells—an embryonic cell population integral to the formation of facial features, heart structures, nervous systems, and pigmentation. Coupled with Theis’s computational neuroscience expertise and the deep learning acumen of doctoral researcher Weixu Wang, RegVelo represents a seamless integration of experimental precision and algorithmic innovation.

To rigorously evaluate RegVelo’s predictive capacity, the team applied their model to multiple biological contexts, including classic developmental archetypes like the cell cycle, hematopoiesis (blood cell formation), and pancreatic organogenesis. The most exhaustive application targeted zebrafish neural crest cells, chosen for their versatility and well-characterized developmental fate decisions. Crucially, RegVelo pinpointed the transcription factor tfec as an early, pivotal regulator of pigment cell development, a finding concordant with prior knowledge. More strikingly, the model identified elf1, a previously unrecognized regulator, as a key driver in melanocyte differentiation. These insights transcended computational predictions, as subsequent CRISPR/Cas9 knockouts and single-cell Perturb-seq experiments validated these regulatory roles, underscoring RegVelo’s robustness.

“Development has too often been depicted as static snapshots of cellular states,” reflects Sauka-Spengler. “RegVelo changes this narrative by capturing the continuous decision-making process—a cell’s journey through transient regulatory landscapes that dictate its fate. This temporal and mechanistic clarity paves the way to decipher how switching a single gene on or off can reroute a developmental path entirely.” The ability to simulate genetic perturbations in silico anticipates transformative impacts on experimental design, enabling researchers to prioritize hypotheses with real-time, predictive insights.

From a translational perspective, RegVelo heralds a significant leap toward the realization of ‘virtual cell’ models that can forecast cell behavior under genetic manipulations with unprecedented accuracy. Such models hold immense promise for disease modeling, especially in understanding how disruptions in regulatory networks contribute to developmental disorders, oncogenesis, or regenerative processes. The framework’s predictive precision opens opportunities for identifying novel therapeutic targets by revealing hidden nodes within gene regulatory landscapes that might be amenable to intervention.

Furthermore, the integration of gene regulatory networks with dynamic velocity modeling exemplifies a blueprint for multi-modal computational biology frameworks, harmonizing discrete data types into cohesive, interpretable models. This approach exemplifies the power of AI to detect and model emergent properties of biological systems that elude traditional analysis, guiding science toward predictive and prescriptive biology rather than merely descriptive.

Looking ahead, Prof. Theis envisions RegVelo as a cornerstone technology for the next generation of stem cell research and cellular engineering. “By simulating and validating gene regulatory perturbations, we are edging closer to being able to rationally engineer cell fates, steering naïve cells into desired lineages with medical relevance. This has profound implications for cell therapy and regenerative medicine, where controlling differentiation pathways is essential for successful outcomes.”

In sum, the advent of RegVelo signifies a critical inflection point in developmental biology research, combining the temporal dimension of single-cell trajectories with the mechanistic context of regulatory networks within a unified AI framework. This synergy not only augments our fundamental understanding of cellular decision-making but also accelerates the translation of single-cell omics data into actionable biological insights, potentially revolutionizing therapeutic strategies across diverse biomedical fields.

Subject of Research: Animals

Article Title: RegVelo: gene-regulatory-informed dynamics of single cells

News Publication Date: 11-May-2026

Web References:

• DOI: 10.1016/j.cell.2026.04.022
• Stowers Institute for Medical Research: www.stowers.org
• Helmholtz Munich: www.helmholtz-munich.de/en
• University of Oxford MRC Weatherall Institute of Molecular Medicine: imm.ox.ac.uk
• Radcliffe Department of Medicine: www.rdm.ox.ac.uk

References:
Sauka-Spengler T. et al., “RegVelo: gene-regulatory-informed dynamics of single cells,” Cell, 2026, DOI: 10.1016/j.cell.2026.04.022.

Image Credits: Stowers Institute for Medical Research

Keywords

Gene regulatory networks, single-cell RNA velocity, developmental biology, AI modeling, computational biology, neural crest cells, zebrafish, transcriptomics, CRISPR/Cas9, Perturb-seq, cell fate prediction, virtual cell models

sc-SPLASH represents a transformative shift in the analytical paradigm of scRNA-seq data. Instead of relying on prior knowledge encoded in existing genomic references, this method harnesses barcoded data in a manner that identifies novel transcriptomic features independent of alignment. This is especially crucial when analyzing species with incomplete or missing reference genomes, such as the sponge Spongilla and the tunicate Ciona, organisms that occupy pivotal evolutionary positions but have been historically genomically understudied. The ability to uncover previously unknown genomic elements and transcript variants in such organisms opens new avenues in evolutionary biology and functional genomics.

Central to the sc-SPLASH framework is its BKC submodule, a component engineered to optimize preprocessing of barcoded sequencing data. Preprocessing of such data—especially unique molecular identifiers (UMIs) which are critical for eliminating amplification bias and counting transcripts accurately—can be a computational bottleneck. Remarkably, BKC has been demonstrated to operate approximately 50 times faster than the commonly used UMI-tools pipeline, commonly considered an industry standard. This incredible improvement in speed and efficiency will enable researchers to process larger datasets more swiftly, accelerating discovery cycles and reducing computational resource demands.

The reference-free aspect of sc-SPLASH leverages sophisticated statistical models that detect transcriptomic variation at the barcode level before any alignment steps take place. This approach allows for an unbiased detection of features such as secreted repeat proteins, which can be obscured or entirely missed when relying solely on reference-guided methods. In their analyses, the authors discovered immune-like cells in both Spongilla and Ciona exhibiting expression of secreted repeat proteins that were notably absent from existing reference annotations. These findings suggest a rich layer of functional complexity previously masked by conventional analytical pipelines.

From a technical perspective, sc-SPLASH integrates advanced algorithms capable of handling the massive scale and complexity of barcode-laden single-cell datasets. The method efficiently resolves barcode errors and PCR duplicates, crucial for ensuring data integrity and reducing false positives. Moreover, it captures subtle transcriptomic features that could be lost in noise or mistaken for artifacts in other methods. This is achieved through a carefully designed statistics-first workflow that prioritizes authentic biological signal from the outset.

The implications of deploying sc-SPLASH extend beyond single-cell research into spatial transcriptomics, where gene expression is mapped within tissue architecture. By applying a reference-free analysis, spatial transcriptomic studies can unveil novel cell types, states, or spatially restricted transcript variants that defy current genomic annotations. This capability enhances our understanding of tissue complexity and cellular interactions, particularly in non-model organisms and novel experimental contexts.

Notably, the authors emphasize that sc-SPLASH empowers open-ended discovery. Without preconceptions imposed by incomplete or biased reference genomes, researchers are freer to uncover unanticipated biology. This is particularly relevant in evolutionary and environmental biology, where many species lack comprehensive genomic resources. By revealing new classes of proteins and transcript variants, even in well-studied organisms, this tool opens the door for novel hypotheses about cellular function and evolution.

The dramatic speed improvement from the BKC submodule means that large-scale studies involving hundreds of thousands to millions of cells—the scale at which many cutting-edge single-cell atlases operate—become feasible within reasonable timeframes and computing budgets. This gains paramount importance as scRNA-seq experiments increasingly push towards whole-organism or multi-organ datasets that generate vast volumes of barcoded reads.

In addition, sc-SPLASH’s focus on barcoded data acknowledges the modern realities of single-cell sequencing. Barcodes, including cell and molecule identifiers, are crucial for deconvoluting complex data but also introduce noise and error. The methodological sophistication of sc-SPLASH in preprocessing these barcodes ensures that downstream biological inference is not compromised, establishing a new benchmark for data quality and reliability in the field.

Perhaps most exciting, the application of sc-SPLASH to uncover immune-like cells expressing secreted repeat proteins in relatively unexplored metazoans like Spongilla and Ciona underscores how new computational tools can revive biological inquiry into basal animal lineages. These discoveries are not just esoteric; they have the potential to reshape our understanding of immune system evolution, the diversification of repeat protein functions, and possibly biotechnological applications where novel repeats may serve as scaffolds or bioactive molecules.

As single-cell sequencing technologies evolve and expand, tools like sc-SPLASH are essential to harness the full information content inherent in these high-dimensional datasets. By eliminating reliance on incomplete references, reducing computational costs, and focusing on intrinsic statistical properties of barcoded data, this method sets a new standard for exploratory transcriptomics. It invites a move away from the constraints of known gene catalogs towards a more holistic, unbiased exploration of cellular identity and function.

Furthermore, the integration of sc-SPLASH with existing computational pipelines will be straightforward for many laboratories. Its design emphasizes compatibility with raw barcoded reads and modularity in preprocessing, meaning that even well-established workflows can benefit from its optimized speed and statistical rigor. This accessibility will likely accelerate its adoption and catalyze discoveries across diverse biological fields.

The publication of this method at a time when single-cell and spatial transcriptomics are becoming foundational across basic, translational, and clinical research domains ensures its broad relevance. From decoding microbial communities to exploring human disease heterogeneity, reference-free discovery methods fill a critical gap by overcoming the limitations of incomplete or absent genomic references.

In conclusion, sc-SPLASH represents a leap forward in the analytical toolkit available to the life sciences community. By enabling reference-free, statistics-first interrogation of barcoded single-cell and spatial transcriptomic data, it opens new vistas for discovery in both model and non-model organisms alike. The ability to detect hidden transcriptomic complexity with unprecedented speed and accuracy heralds a new age of functional genomics that prizes unbiased exploration as much as hypothesis-driven inquiry.

This innovative approach not only enhances our capacity to characterize novel cell types and molecular players but also challenges the community to reconsider the dominance of reference-based paradigms in the era of big data biology. The future of single-cell research will undoubtedly be shaped by such advances that allow the data to speak for itself, unshackled from the confines of prior knowledge.

Subject of Research: Reference-free, barcoded single-cell RNA sequencing and spatial transcriptomics data analysis.

Article Title: Reference-free discovery with barcoded single-cell sequencing.

Article References:
Dehghannasiri, R., Kokot, M., Starr, A.L. et al. Reference-free discovery with barcoded single-cell sequencing. Nat Biotechnol (2026). https://doi.org/10.1038/s41587-026-03084-6

Image Credits: AI Generated

DOI: https://doi.org/10.1038/s41587-026-03084-6

Single-cell sequencing technology has revolutionized the way we analyze cellular diversity. Traditional bulk sequencing methods averaged the molecular profiles of millions of cells, masking the intricate variations that exist at the individual cell level. The shift toward single-cell techniques has unveiled a new world of information, revealing distinct cellular states that play critical roles in health and disease. However, the challenge remains in accurately annotating the vast amounts of data generated from these powerful technologies. This is where the new approach by Cultrera di Montesano and colleagues comes into play.

The authors of the study detail a model that employs hierarchical cross-entropy loss for single-cell annotation. The key innovation here lies in how the model structures its learning pathway. Traditional models often rely on simple loss functions, which treat each misclassification with equal weight. In contrast, the hierarchical approach allows the model to prioritize certain classifications based on their biological significance. This is particularly important in complex biological systems where some cell types may be more relevant to disease mechanisms than others.

This novel hierarchical framework enables the model to learn not just from individual cell characteristics but also from the relationships between various cell types. By recognizing these hierarchical relationships, the model can better navigate the vast landscape of single-cell transcriptomic data. For instance, when annotating a dataset that includes both immune and epithelial cells, the model understands that immune cells can be further sub-categorized into distinct populations such as T cells, B cells, and macrophages. The innovative use of cross-entropy loss within this structure provides a more nuanced training experience, leading to improved performance on validation datasets.

The implementation of this model also represents a significant departure from more conventional methods. While past approaches have often relied on linear relationships in the data, the hierarchical model accepts the complexity of biological interactions, recognizing that cell types do not exist in isolation. This is a crucial aspect that many traditional approaches have overlooked, limiting their application in real-world scenarios where cellular interplay is vital for understanding disease progression and treatment responses.

In their article, Cultrera di Montesano and colleagues provide an extensive evaluation of their model on various datasets. Their results indicate that the hierarchical cross-entropy loss approach consistently outperforms classic single-cell annotation methods across multiple benchmarks. This performance boost suggests that the model not only achieves a higher accuracy rate but also enhances the discernment of subtle differences among closely related cell types. This could prove indispensable in fields such as oncology, where distinguishing between cancerous and non-cancerous cells at a single-cell level is crucial for effective treatment strategies.

One of the key takeaways from this research is the importance of model interpretability. Given the intricacies of biological data, it is imperative that researchers not only obtain accurate annotations but also understand the rationale behind these classifications. The hierarchical model offers insights into the decision-making process, allowing researchers to trace the lineage of their annotations back to specific characteristics within the data. This interpretability can foster greater trust among researchers in the annotations produced, encouraging broader adoption of advanced computational methods in biological research.

Furthermore, the implications of this work extend beyond just theoretical advancements. As single-cell technologies continue to evolve, so do the datasets that accompany them. With the proliferation of large-scale scRNA-seq datasets, researchers face challenges not only in computational power to analyze these datasets but also in developing methodologies that can keep pace with the data’s growth. The hierarchical model stands as a form of optimism in tackling these challenges, providing a scalable solution to single-cell analysis.

Another vital aspect of the study is its potential application in personalized medicine. Single-cell annotations are critical in delineating patient-specific profiles that inform treatment options, especially in precision oncology. By leveraging the improved accuracy provided by the hierarchical model, clinicians may be better equipped to identify therapeutic targets and monitor disease progression on an individual level. Such capabilities could revolutionize how treatments are tailored to patients, leading to better outcomes.

Moreover, as the field of machine learning continues to integrate with biological sciences, this study illustrates the importance of interdisciplinary approaches. The collaboration between computational scientists and biologists is paramount for translating complex algorithms into practical solutions for real-world biological questions. This work by Cultrera di Montesano et al. exemplifies how such collaborations can yield innovative methodologies that push the boundaries of our current understanding in biology.

Enabling such advancements also requires a commitment to open science. The document highlights that the authors have made their code and methodologies available to the research community, ensuring that other scientists can build upon their work. This transparent approach fosters a collaborative environment where ideas can be shared and developed further, which is critical for collective advancement in understanding cellular biology.

In conclusion, the research led by Cultrera di Montesano et al. marks a significant stride in the realm of single-cell annotation. By introducing a hierarchical framework utilizing cross-entropy loss, the authors have set a new standard for accuracy and interpretability in this crucial area. With continued advancements in technology and methodology, the future of single-cell analysis looks promising, opening doors to discoveries that could have profound implications in health, disease, and personalized medicine. The impact of this research will likely echo across various domains of biological science, influencing future studies and methodologies in how we understand and manipulate biological systems.

As scientists continue to grapple with the complexities of cellular biology, the insights derived from this research will undoubtedly serve as a springboard for future exploration in this ever-evolving field. The hierarchical cross-entropy loss framework is not just a technical advancement; it represents a paradigm shift in how we perceive and engage with the biological intricacies at the single-cell level.

Subject of Research: Single-cell annotation models and hierarchical cross-entropy loss in computational biology.

Article Title: Improving atlas-scale single-cell annotation models with hierarchical cross-entropy loss.

Article References: Cultrera di Montesano, S., D’Ascenzo, D., Raghavan, S. et al. Improving atlas-scale single-cell annotation models with hierarchical cross-entropy loss. Nat Comput Sci (2026). https://doi.org/10.1038/s43588-025-00945-z

Image Credits: AI Generated

DOI: https://doi.org/10.1038/s43588-025-00945-z

Keywords: single-cell analysis, hierarchical cross-entropy loss, cell annotation, computational biology, precision medicine, machine learning, biological data.

COSIME, or Cooperative Multi-view Integration with a Scalable and Interpretable Model Explainer, harnesses the power of deep learning to navigate the complexities of biological data integration. This model utilizes a backpropagation technique grounded in optimal transport algorithms, which elegantly facilitates the extraction of latent features from diverse data views. By leveraging these sophisticated techniques, COSIME aims not only to predict disease phenotypes effectively but also to unveil the subtle, yet critical, interactions among biological features. This dual capability is essential for gaining a holistic understanding of diseases that manifest through multifactorial processes.

The growing challenge in biological research is the integration and analysis of multi-omics data, which can include single-cell transcriptomics, spatial transcriptomics, epigenomics, and metabolomics. Each of these data types provides a unique perspective on cellular processes, but their combination often reveals interactions that cannot be understood through a singular lens. COSIME addresses these challenges head-on by employing a robust model that synergizes these diverse data types, creating a comprehensive view of the biological landscape. Thus, COSIME opens avenues for research that evaluate intricate feature interactions across different biological dimensions.

What sets COSIME apart from traditional models is its incorporation of Monte Carlo sampling techniques, which foster interpretable assessments at both the feature importance level and the pairwise interaction level. This feature is particularly significant, as it allows researchers to derive meaningful insights from complex datasets without the risk of oversimplifying the relationships at play. By providing a nuanced interpretation of the data, COSIME enhances our understanding of how different biological features might interrelate, ultimately leading to more informed hypotheses and research directions.

To test the efficacy of COSIME, researchers employed it across a variety of datasets, ranging from simulated environments to real-world applications involving Alzheimer’s disease-related phenotypes. The model proved to be a watershed moment in the predictive accuracy of disease characteristics, eclipsing existing methodologies in its performance. The enhanced prediction accuracy is significant not only for theoretical research but also for clinical applications where accurate phenotype prediction could profoundly affect patient care and treatment outcomes.

For instance, one of the critical discoveries made using COSIME was the identification of synergistic interactions between astrocyte and microglia genes related to Alzheimer’s disease. This revelation holds practical implications for neurobiological understanding, suggesting that these particular gene interactions may localize to specific areas within the brain, such as the edges of the middle temporal gyrus. Such insights are invaluable, shedding light on disease mechanisms that were previously underexplored or entirely overlooked due to data siloing.

Recognizing the broad applicability and the need for accessible tools in scientific research, the creators of COSIME made it publicly available as an open-source resource. This transparency not only encourages wider adoption among researchers in diverse fields but also fosters a collaborative environment wherein users can contribute to and improve the model. An open-source approach democratizes access to advanced analytical techniques, promoting rigorous scientific inquiry across disciplines.

Moreover, the introduction of COSIME highlights a growing trend within computational biology that emphasizes interpretability. While machine learning models have historically been viewed as “black boxes”, new strategies are emerging to ensure that the relationships discovered by these models are understandable to biologists. This shift is crucial as it empowers researchers to validate findings within their biological contexts and integrate them meaningfully into their ongoing research.

The implications of COSIME extend beyond Alzheimer’s disease. As the model demonstrates versatility with various types of omics data, it stands to redefine how we approach various complex diseases. From cancer biology to metabolic disorders, the ability to holistically integrate multiple data types allows for the possibility of uncovering novel biomarkers and therapeutic targets that could have significant implications for clinical practice.

Additionally, the continuous evolution of computational techniques suggests that we are only beginning to scratch the surface of what is possible with multi-omics data integration. As new datasets become available and computational power increases, models akin to COSIME will likely become instrumental in shaping future biological research. By bridging gaps between disparate data types and providing robust interpretive frameworks, such models can guide the next generation of discoveries in molecular biology and medicine.

Finally, as we move toward a future that increasingly relies on personalized medicine and targeted therapies, tools like COSIME will be paramount in guiding research directions. The ability to accurately predict disease phenotypes and elucidate underlying biological interactions will not only enhance our understanding of complex diseases but also directly inform treatment strategies that can be tailored to individual patients. This personalized approach, powered by multi-omics data integration, holds startling potential for improving patient outcomes and advancing the field of medicine as a whole.

Subject of Research: Multi-omics integration for understanding complex biological systems and disease phenotypes.

Article Title: Cooperative multi-view integration with a scalable and interpretable model explainer.

Article References:

Choi, J.J., Cohen Kalafut, N., Gruenloh, T. et al. Cooperative multi-view integration with a scalable and interpretable model explainer.
Nat Mach Intell (2025). https://doi.org/10.1038/s42256-025-01111-w

Image Credits: AI Generated

DOI: 10.1038/s42256-025-01111-w

Keywords: Multi-omics, Disease phenotypes, COSIME, Data integration, Alzheimer’s disease, Machine learning, Interpretability, Biomarkers, Personalized medicine.