Results show that signed dendritic gating supports competitive clustering with a parameter-inspectable decision function, while indicating that decision-relevant information is distributed across the latent space.
Abstract
Deep clustering models for single-cell RNA sequencing often assign cells through latent or centroid-based mechanisms that are difficult to inspect. We introduce scDNM-VAE (single-cell Dendritic Neuron Model Variational Autoencoder), a deep clustering framework that combines a variational autoencoder with a dendritic neuron-inspired head. Cluster assignments are governed by learnable signed synaptic weights and thresholds: the weight sign determines the direction of a gate's response to a latent coordinate, its magnitude controls steepness, and the weight-threshold pair determines the transition location. The trained clustering function can therefore be inspected directly without fitting a post-hoc explanation model. We benchmark scDNM-VAE on four datasets spanning immune, cortical, cardiac, and hematopoietic cells against scVI followed by KMeans and an MLP-DEC ablation. scDNM-VAE performs better than scVI on PBMC3k, comparably on the Human Heart Cell Atlas and Paul15, and worse on Zeisel, while producing biologically coherent marker-gene signatures. Ablating each cluster's three highest-magnitude synaptic dimensions causes numerically greater reassignment than random-dimension ablation across all datasets, but the margins are modest and negligible on Zeisel. These results show that signed dendritic gating supports competitive clustering with a parameter-inspectable decision function, while indicating that decision-relevant information is distributed across the latent space.
Single-cell RNA sequencing (scRNA-seq) provides a novel perspective to explore cellular biology at the single-cell resolution. Single-cell clustering is a crucial step to reveal cell types and the corresponding biological functions. However, when dealing with the high dimensionality and complexity of scRNA-seq data, existing deep models fail to comprehensively capture the intrinsic attribute information and structural relationships within the data. In this study, we propose a novel single-cell deep clustering model named scDFVA. The proposed scDFVA consists of a variational graph attention autoencoder (AE), a zero-inflated negative binomial (ZINB) based AE, and a self-supervised clustering. To better simulate sparse and zero-inflated scRNA-seq data, we incorporate the ZINB model into the AE. The variational graph attention AE is introduced to learn the cell structure information. scDFVA achieves representation learning within a joint framework comprising a ZINB-based AE and a variational graph attention AE, effectively fusing gene expression and cell structure information. Furthermore, scDFVA performs self-supervised clustering training on the latent fusion representations of cells to achieve mutual supervision between representation learning and clustering. Experiments indicated that scDFVA outperformed several other competing methods, demonstrating that our method is beneficial in single-cell clustering.
Ge Zhang, Maohua Qin, Xuye Kou et al.· J. Comput. Biol.· 0 citations
Per-dataset analysis reveals three reproducible regimes: probabilistic variational autoencoder variants help on the smallest datasets, deep autoencoders win on mid-scale data with multi-batch or many-type structure, and classical PCA pipelines remain competitive when linear projection already captures the dominant variation.
Phong T. Nguyen, T. Vu, T. Nguyen et al.· arXiv.org· 0 citations
Integrating unpaired single-cell RNA-seq and ATAC-seq data remains challenging due to the profound differences in data sparsity and noise between them. This paper presents a novel discrete representation learning based method for effective and robust cross-omics cell-type annotation, which is called scCoA-VQA --- the abbreviation of single-cell Cross-omics Annotation via Vector-Quantized Autoencoders. scCoA-VQA employs omics-specific vector-quantized autoencoders to construct stable discrete latent spaces, and aligns RNA and ATAC representations via an anchor-constrained autoencoder constrained by biologically meaningful intra- and inter-omics anchors. A two-phase label transfer strategy is proposed to achieve accurate label transfer by combining inter-omics propagation with intra-ATAC refinement. Extensive experiments on both synthetic and real-world datasets show that scCoA-VQA consistently outperforms existing methods in accuracy, F1-Macro, and robustness to extreme sparsity. Further biological analysis demonstrate that scCoA-VQA can accurately capture the meaningful regulatory structure, including transitional Naive–Effector T-cell states and Naive-like versus Memory-like epigenetic subpopulations in CD4 TCM cells. These results indicate that discrete, anchor-constrained latent modeling provides a powerful and biologically coherent solution for unpaired single-cell multi-omics integration. Source code of this work is available at https://github.com/penghan-ph/scCoA-VQA/.
Han Peng, Wuchao Liu, Yifang Cai et al.· Proceedings of the 32nd ACM...· 0 citations
ACSCeND offers a robust, interpretable approach to profiling CSC dynamics and establishes CSC state as a clinically meaningful, pan-cancer biomarker for guiding stemness-informed therapies by integrating single-cell precision with bulk-level applicability.
Single-cell RNA sequencing (scRNA-seq) captures detailed gene expression profiles at scale, while patch-clamp recordings measure intrinsic neuronal electrophysiological properties. Modeling the relations between these two modalities remains a challenge. Here, we compare how well electrophysiological features can be predicted by traditional transcriptomic cell type classification, representations derived from a foundational model (scGPT) pretrained on large-scale scRNA-seq datasets, ion channel-coding genes, and highly variable genes. Using paired transcriptomic and electrophysiological patch-sequencing data from 495 human neurons from neurosurgical tissue, we find that cluster-level cell type representations consistently outperform highly variable gene selection, ion channel gene selection, and context-enriched scGPT embeddings. Notably, performance varies across model architectures and initializations, and the best results are obtained by combining the outputs of separate cell type and scGPT-based models. Together, these findings suggest that traditional discrete cellular classification is highly effective in predicting physiological features. For maximum performance it can be complemented by pretrained transformer models. Author summary Understanding how a neuron’s genes relate to its electrical properties is a major goal in neuroscience. New technologies now make it possible to measure gene expression in individual cells and, simultaneously, to record how those same cells respond to electrical signals. However, relating these two types of information at the single cell level remains difficult. In this study, we tested whether a modern artificial intelligence model trained on large collections of gene expression data could help connect gene activity to electrical behavior in human brain cells. We compared this approach with simpler strategies, such as using selected sets of genes or grouping cells by their known types. We found that basic cell type descriptions often predicted electrical properties better than more complex gene-based methods alone. The strongest results were achieved by combining cell type knowledge with information from the artificial intelligence model and training them together. These findings suggest that when linking data is in the hundreds, simpler representations outperform large general-purpose AI models in isolation, but the two approaches may be complementary rather than competing.
The framework integrates three modules: dual-reconstruction to fuse attribute-structure information, contrastive learning under label guidance to extract semantic similarities, and deep embedding clustering to enable iterative optimization.
Wenjing Su, Baojuan Qin, Junliang Shang et al.· Interdisciplinary Sciences C...· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
