Biology subjects

Kempynck, N.

Publications and source records attributed to Kempynck, N..

12 recordsLinked to original sources

bioRxiv2026

A subtype-specific gene regulatory program controls differential circuit integration of deep CA3 pyramidal neurons

How closely related neuronal subtypes integrate into different circuits remains poorly understood. Deep and superficial hippocampal CA3 pyramidal neurons (PNs) differ in intrinsic firing and mossy fiber (MF) input at specialized postsynaptic structures termed thorny excrescences (TEs). Combining developmental single-nucleus multiomics with genetic targeting, morphological analysis and electrophysiology, we identify the transcription factor ST18 as a regulator of deep CA3 subtype identity. ST18 marks sparsely thorny and burst-firing neurons, and gene regulatory analyses predict that it represses a superficial-associated program involved in dendritic spine development, synaptic organization, and electrophysiological properties. Postnatal ST18 loss shifts the transcriptional and chromatin landscape of deep CA3 PNs toward a superficial-like state. These molecular changes are accompanied by increased TE elaboration approaching superficial levels, enhanced excitatory synaptic activity, and reduced burst-firing. Together, these findings identify ST18-dependent restriction of an alternative subtype program as a mechanism linking molecular identity to distinct input connectivity patterns and intrinsic firing properties during postnatal circuit maturation.

neuroscience↗

bioRxiv2026

Evolutionary transfer learning enables organism-wide inference of mammalian enhancer landscapes

Understanding and modeling how a single human genome concurrently encodes gene regulatory programs for thousands of cell types remains a central challenge in genomics and machine learning. Most human cell types emerge during embryonic, fetal, and pediatric development which are inaccessible to comprehensive molecular profiling. To circumvent this, we hypothesized that the mismatch in evolutionary rates between cis-acting enhancers that modulate gene expression (fast) and the trans-acting regulatory factors that specify cell types (slow) creates an opportunity for evolutionary transfer learning. Specifically, models trained to predict cell type-specific enhancers in one species should generalize to the orthologous cell types and enhancers of related species. To test this, we generated a single-cell atlas of chromatin accessibility spanning mouse embryonic day 10 (E10) to birth (P0). Using combinatorial indexing1, we profiled 3.9 million nuclei from 36 staged embryos, resolving genome-wide accessibility in 36 cell classes and 140 cell types. We then trained a series of multi-output deep learning models (CREsted2), each addressing limitations of the preceding approach, towards the goal of genome-wide prediction of distal enhancers across major developmental lineages. An evolution-naive model achieved strong performance on heldout peaks, but exhibited two failure modes during genome-wide inference: overprediction at tandem repeats and conflation of promoter and distal enhancer grammars. An evolution-aware model resolved these by regrouping accessible regions based on their retention and functional coherence across mammalian evolution, but failed to generalize across species. Finally, an evolution-augmented model, STEAM (Synteny-aware Transfer learning for Enhancer Activity Modeling), incorporated enhancer orthologs from 241 mammalian genomes (Zoonomia3) in a synteny-supervised manner. This increased the effective data scale by as much as 195-fold, markedly improving generalization across mammals despite greater label noise. We applied STEAM to the genome-wide inference of cell class-specific distal developmental enhancers in humans, mice (HumMus) and 239 additional mammals3 (BabaGanoush), i.e. 32 x 241 = 7,712 genome-wide distal enhancer prediction tracks. Together, our results unify advances in single-cell profiling, deep learning, and comparative genomics into a framework for the evolutionary transfer learning of noncoding regulatory grammars. More broadly, our work supports the view that model organisms and evolutionarily diverse genomes are indispensable resources for accelerating and enhancing the AI-enabled exploration of human biology. NoteAn interactive version of this preprint, together with count matrices, CREsted models, prediction tracks, code and reproducible figures, is available at this link (ref 4).

genomics↗

bioRxiv2026

A Cross-Species Enhancer-AAV Toolkit for Cell Type-Specific Targeting Across the Basal Ganglia

The mammalian basal ganglia (BG) orchestrate motor, cognitive, and affective functions, yet cell type-specific genetic access remains limited, especially beyond rodents. Key structures implicated in movement and psychiatric disorders, including pallidum, subthalamic nucleus, and dopaminergic midbrain, lack scalable tools for cross-species targeting. Here, we present a comprehensive enhancer-AAV library enabling selective labeling and manipulation of major BG neuronal populations: striatal projection neuron subtypes, pallidal and subthalamic neurons, and midbrain dopaminergic and GABAergic populations. Using an evolutionarily informed discovery pipeline, we identified enhancers targeting canonical, non-canonical, and disease-relevant cell types, with validation demonstrating robust cross-species conservation of specificity between mouse and macaque. Computational modeling revealed sequence features predictive of in vivo performance, including motif grammar, chromatin accessibility, and evolutionary conservation, and identified distinct regulatory architectures across glial, projection, and interneuron lineages. This work establishes a comprehensive cross-species viral toolkit for the BG, unlocking previously inaccessible cell types for circuit dissection.

neuroscience↗

bioRxiv2026

A consensus spinal cord cell type atlas across mouse, macaque, and human

The spinal cord contains evolutionarily conserved cell types critical for motor function, sensory processing, and autonomic regulation, many of which are implicated in diverse neurological diseases and injuries. Yet the field lacks a comprehensive molecular characterization of cellular diversity in human, macaque, and mouse spinal cord. Here, we present a unified, cross-species cell type atlas based on the integration of single-nucleus gene expression, chromatin accessibility, and spatial transcriptomic data from segments within cervical, thoracic, lumbar, and sacral regions, including motor neurons (MNs) sampled across the entire rostro-caudal axis of the macaque spinal cord. Leveraging the spatial distributions of our molecularly defined cell types, we generated a cell type-guided anatomical map of spinal cord laminae and nuclei. We identified both conserved and species-specific cellular features, including gene expression patterns across distinct MN subtypes in the primate spinal cord. Cross-species cis-regulatory analysis and deep learning sequence models dissected the enhancer logic underlying viral targeting, uncovering conserved transcription factor grammar encoding cellular identity. Together, these results establish a unifying molecular and anatomical taxonomy of spinal cord cell types across species.

neuroscience↗

bioRxiv2025

Circuit specific specialization of human basal ganglia astrocytes

Astrocytes shape synapses and circuits, yet human basal ganglia astrocyte diversity is incompletely defined. We built a multimodal atlas by integrating single-nucleus RNA-sequencing and chromatin accessibility with DNA methylation, 3D chromatin conformation, and spatial transcriptomics, then mapped basal ganglia programs onto a whole-brain reference. Astrocytes segregated into three anatomical subgroups spanning striatal gray matter, extra-striatal gray matter, and white matter, with subgroup-biased neurotransmitter transporters and synapse-associated programs consistent with differences in dominant afferent input. Within striatum, dorsal and ventral astrocyte populations aligned with distinct microcircuits and were conserved in nonhuman primates. A deep learning sequence model identified subgroup-associated enhancer code and, when benchmarked against published enhancer-AAV datasets, supported the design of candidate viral tools to target basal ganglia astrocyte programs in vivo. Together, these data define major axes of human astrocyte specialization and provide a framework for cell type-specific dissection of basal ganglia function.

neuroscience↗

bioRxiv2025

Cross-species consensus atlas of the primate basal ganglia

The basal ganglia (BG) are conserved brain regions essential for motor control, learning, emotion, and cognition, and are implicated in neurological and psychiatric disease. Yet a unified cross-species taxonomy of BG cell types is lacking, limiting translation of BG circuit mechanisms, interpretation of human genetic risk, and development of cell type-targeted tools. We present a multiomic consensus atlas of 1.8 million nuclei from human, macaque, and marmoset spanning eight BG structures. Integrating cross-species gene expression, open chromatin, and spatial profiling enables definition of conserved and divergent cell types. Alignment to existing mouse and human atlases identifies 61 homologous cell types conserved over 80 million years. We identify a STRd D2 StrioMat Hybrid medium spiny neuron (MSN) type with molecular, electrophysiological, and morphological features that clarify hybrid MSN identities. Comparative cis-regulatory analysis reveals conserved sequence grammars that encode cell identity and inform viral targeting strategies, providing a foundational resource for BG evolution, function, and disease.

neuroscience↗

bioRxiv2025

GAME: Genomic API for Model Evaluation

The rapid expansion of genomics datasets and the application of machine learning has produced sequence-to-activity genomics models with ever-expanding capabilities. However, benchmarking these models on practical applications has been challenging because individual projects evaluate their models in ad hoc ways, and there is substantial heterogeneity of both model architectures and benchmarking tasks. To address this challenge, we have created GAME, a system for large-scale, community-led standardized model benchmarking on user-defined evaluation tasks. We borrow concepts from the Application Programming Interface (API) paradigm to allow for seamless communication between pre-trained models and benchmarking tasks, ensuring consistent evaluation protocols. Because all models and benchmarks are inherently compatible in this framework, the continual addition of new models and new benchmarks is easy. We also developed a Matcher module powered by a large language model (LLM) to automate ambiguous task alignment between benchmarks and models. Containerization of these modules enhances reproducibility and facilitates the deployment of models and benchmarks across computing platforms. By focusing on predicting underlying biochemical phenomena (e.g. gene expression, open chromatin, DNA binding), we ensure that tasks remain technology-independent. We provide examples of benchmarks and models implementing this framework, and anticipate that the community will contribute their own, leading to an ever-expanding and evolving set of models and evaluation tasks. This resource will accelerate genomics research by illuminating the best models for a given task, motivating novel functional genomic benchmarks, and providing a more nuanced understanding of model abilities.

bioinformatics↗

bioRxiv2025

HyDrop v2: Scalable atlas construction for training sequence-to-function models

Deciphering cis-regulatory logic underlying cell type identity is a fundamental question in biology. Single-cell chromatin accessibility (scATAC-seq) data has enabled training of sequence-to-function deep learning models allowing decoding of enhancer logic and design of synthetic enhancers. Training such models requires large amounts of high-quality training data across species, organs, development, aging, and disease. To facilitate the cost-effective generation of large scATAC-seq atlases for model training, we developed a new version of the open-source microfluidic system HyDrop with increased sensitivity and scale: HyDrop v2. We generated HyDrop v2 atlases for the mouse cortex and Drosophila embryo development and compared them to atlases generated on commercial platforms. HyDrop v2 data integrates seamlessly with commercially available chromatin accessibility methods (10x Genomics). Differentially accessible regions and motif enrichment across cell types are equivalent between HyDrop-v2 and 10x atlases. Sequence-to-function models trained on either atlas are comparable as well in terms of enhancer predictions, sequence explainability, and transcription factor footprinting. By offering accessible data generation, enhancer models trained on HyDrop-v2 and mixed atlases can contribute to unraveling cell-type specific regulatory elements in health and disease.

bioinformatics↗

bioRxiv2025

CREsted: modeling genomic and synthetic cell type-specific enhancers across tissues and species

Sequence-based deep learning models have become the state of the art for the analysis of the genomic regulatory code. Particularly for transcriptional enhancers, deep learning models excel at deciphering sequence features and grammar that underlie their spatiotemporal activity. To enable end-to-end enhancer modeling and design, we developed a software and modeling package, called CREsted. It combines preprocessing starting from single-cell ATAC-seq data; modeling with a choice of several architectures for training classification and regression models on either topics or pseudobulk peak heights; sequence design using multiple strategies; and downstream analysis through a collection of tools to locate transcription factor (TF) binding sites, infer the effect of a TF (activating or repressing) on enhancer accessibility, decipher enhancer grammar, and score gene loci. We demonstrate CREsted using a mouse cortex model that we validate using the BICCN collection of in vivo validated mouse brain enhancers. Classical enhancers in immune cells, including the IFNB1 enhanceosome are revisited using a PBMC model, and we assess the accuracy of TF binding site predictions with ChIP-seq. Additionally, we use CREsted to compare mesenchymal-like cancer cell states between tumor types; and we investigate different fine-tuning strategies of Borzoi within CREsted, comparing their performance and explainability with CREsted models trained from scratch. Finally, we train a CREsted model on a scATAC-seq atlas of zebrafish development and use this to design and in vivo validate cell type-specific synthetic enhancers in three tissues. For varying datasets, we demonstrate that CREsted facilitates efficient training and analyses, enabling scrutinization of the enhancer logic and design of synthetic enhancers across tissues and species. CREsted is available at https://crested.readthedocs.io.

genomics↗

bioRxiv2025

The evolution of gene regulation in mammalian cerebellum development

Gene regulatory changes are considered major drivers of evolutionary innovations, including the cerebellums expansion during human evolution, yet they remain largely unexplored. In this study, we combined single-nucleus measurements of gene expression and chromatin accessibility from six mammals (human, bonobo, macaque, marmoset, mouse, and opossum) to uncover conserved and diverged regulatory networks in cerebellum development. We identified core regulators of cell identity and developed sequence-based models that revealed conserved regulatory codes. By predicting chromatin accessibility across 240 mammalian species, we reconstructed the evolutionary histories of human cis-regulatory elements, identifying sets associated with positive selection and gene expression changes, including the recent gain of THRB expression in cerebellar progenitor cells. Collectively, our work reveals the shared and mammalian lineage-specific regulatory programs governing cerebellum development.

evolutionary biology↗

bioRxiv2024

Evaluating Methods for the Prediction of Cell Type-Specific Enhancers in the Mammalian Cortex

Identifying cell type-specific enhancers in the brain is critical to building genetic tools for investigating the mammalian brain. Computational methods for functional enhancer prediction have been proposed and validated in the fruit fly and not yet the mammalian brain. We organized the Brain Initiative Cell Census Network (BICCN) Challenge: Predicting Functional Cell Type-Specific Enhancers from Cross-Species Multi-Omics to assess machine learning and feature-based methods designed to nominate enhancer DNA sequences to target cell types in the mouse cortex. Methods were evaluated based on in vivo validation data from hundreds of cortical cell type-specific enhancers that were previously packaged into individual AAV vectors and retro-orbitally injected into mice. We find that open chromatin was a key predictor of functional enhancers, and sequence models improved prediction of non-functional enhancers that can be deprioritized as opposed to pursued for in vivo testing. Sequence models also identified cell type-specific transcription factor codes that can guide designs of in silico enhancers. This community challenge establishes a benchmark for enhancer prioritization algorithms and reveals computational approaches and molecular information that are crucial for identifying functional enhancers in mammalian cortical cell types. The results of this challenge bring us closer to understanding the complex gene regulatory landscape of the mammalian cortex and to designing more efficient genetic tools to target cortical cell types.

genomics↗

bioRxiv2024

Enhancer-driven cell type comparison reveals similarities between the mammalian and bird pallium

Combinations of transcription factors govern the identity of cell types, which is reflected by enhancer codes in cis-regulatory genomic regions. Cell type-specific enhancer codes at nucleotide-level resolution have not yet been characterized for the mammalian neocortex. It is currently unknown whether these codes are conserved in other vertebrate brains, and whether they are informative to resolve homology relationships for species that lack a neocortex such as birds. To compare enhancer codes of cell types from the mammalian neocortex with those from the bird pallium, we generated single-cell multiome and spatially-resolved transcriptomics data of the chicken telencephalon. We then trained deep learning models to characterize cell type-specific enhancer codes for the human, mouse, and chicken telencephalon. We devised three metrics that exploit enhancer codes to compare cell types between species. Based on these metrics, non-neuronal and GABAergic cell types show a high degree of regulatory similarity across vertebrates. Proposed homologies between mammalian neocortical and avian pallial excitatory neurons are still debated. Our enhancer code based comparison shows that excitatory neurons of the mammalian neocortex and the avian pallium exhibit a higher degree of divergence than other cell types. In contrast to existing evolutionary models, the mammalian deep layer excitatory neurons are most similar to mesopallial neurons; and mammalian upper layer neurons to hyper- and nidopallial neurons based on their enhancer codes. In addition to characterizing the enhancer codes in the mammalian and avian telencephalon, and revealing unexpected correspondences between cell types of the mammalian neocortex and the chicken pallium, we present generally applicable deep learning approaches to characterize and compare cell types across species via the genomic regulatory code.

genomics↗