Novel cell types and developmental lineages revealed by single-cell RNA-seq analysis of the mouse crista ampullaris

Deletion of the Ebf1, a mouse deafness gene, causes a dramatic increase in hair cells and support cells of the organ of Corti

Following up on our previous observation that early B cell factor (EBF) sites are enriched in open chromatin of the developing sensory epithelium of the mouse cochlea, we investigated the effect of deletion of Ebf1 on inner ear development. We used a Cre driver to delete Ebf1 at the otocyst stage before development of the cochlea. We examined the cochlea at postnatal day (P) 1 and found that the sensory epithelium had doubled in size but the length of the cochlear duct was unaffected. We also found that deletion of Ebf1 led to ectopic sensory patches in the Kölliker's organ. Innervation of the developing organ of Corti was disrupted with no obvious spiral bundles. The ectopic patches were also innervated. All the extra hair cells (HCs) within the sensory epithelium and Kölliker's organ contained mechanoelectrical transduction channels, as indicated by rapid uptake of FM1-43. The excessive numbers of HCs were still present in the adult Ebf1 conditional knockout (cKO) animal. The animals had significantly elevated auditory brainstem response thresholds, suggesting that this gene is essential for hearing development.

Advances in single-cell transcriptomics in animal research

Understanding biological mechanisms is fundamental for improving animal production and health to meet the growing demand for high-quality protein. As an emerging biotechnology, single-cell transcriptomics has been gradually applied in diverse aspects of animal research, offering an effective method to study the gene expression of high-throughput single cells of different tissues/organs in animals. In an unprecedented manner, researchers have identified cell types/subtypes and their marker genes, inferred cellular fate trajectories, and revealed cell‒cell interactions in animals using single-cell transcriptomics. In this paper, we introduce the development of single-cell technology and review the processes, advancements, and applications of single-cell transcriptomics in animal research. We summarize recent efforts using single-cell transcriptomics to obtain a more profound understanding of animal nutrition and health, reproductive performance, genetics, and disease models in different livestock species. Moreover, the practical experience accumulated based on a large number of cases is highlighted to provide a reference for determining key factors (e.g., sample size, cell clustering, and cell type annotation) in single-cell transcriptomics analysis. We also discuss the limitations and outlook of single-cell transcriptomics in the current stage. This paper describes the comprehensive progress of single-cell transcriptomics in animal research, offering novel insights and sustainable advancements in agricultural productivity and animal health.

3D reconstruction of the mouse cochlea from scRNA-seq data suggests morphogen-based principles in apex-to-base specification

In the mammalian auditory system, frequency discrimination depends on numerous morphological and physiological properties of the organ of Corti, which gradually change along the apex-to-base (tonotopic) axis of the organ. For example, the basilar membrane stiffness changes tonotopically, thus affecting the tuning properties of individual hair cells. At the molecular level, those frequency-specific characteristics are mirrored by gene expression gradients; however, the molecular mechanisms controlling tonotopic gene expression in the mouse cochlea remain elusive. Through analyzing single-cell RNA sequencing (scRNA-seq) data from E12.5 and E14.5 time points, we predicted that morphogens, rather than a cell division-associated mechanism, confer spatial identity in the extending cochlea. Subsequently, we reconstructed the developing cochlea in 3D space from scRNA-seq data to investigate the molecular pathways mediating positional information. The retinoic acid (RA) and hedgehog pathways were found to form opposing apex-to-base gradients, and functional interrogation using mouse cochlear explants suggested that both pathways jointly specify the longitudinal axis.

The biomedical knowledge graph of symptom phenotype in coronary artery plaque: machine learning-based analysis of real-world clinical data

A knowledge graph can effectively showcase the essential characteristics of data and is increasingly emerging as a significant means of integrating information in the field of artificial intelligence. Coronary artery plaque represents a significant etiology of cardiovascular events, posing a diagnostic challenge for clinicians who are confronted with a multitude of nonspecific symptoms. To visualize the hierarchical relationship network graph of the molecular mechanisms underlying plaque properties and symptom phenotypes, patient symptomatology was extracted from electronic health record data from real-world clinical settings. Phenotypic networks were constructed utilizing clinical data and protein‒protein interaction networks. Machine learning techniques, including convolutional neural networks, Dijkstra's algorithm, and gene ontology semantic similarity, were employed to quantify clinical and biological features within the network. The resulting features were then utilized to train a K-nearest neighbor model, yielding 23 symptoms, 41 association rules, and 61 hub genes across the three types of plaques studied, achieving an area under the curve of 92.5%. Weighted correlation network analysis and pathway enrichment were subsequently utilized to identify lipid status-related genes and inflammation-associated pathways that could help explain the differences in plaque properties. To confirm the validity of the network graph model, we conducted coexpression analysis of the hub genes to evaluate their potential diagnostic value. Additionally, we investigated immune cell infiltration, examined the correlations between hub genes and immune cells, and validated the reliability of the identified biological pathways. By integrating clinical data and molecular network information, this biomedical knowledge graph model effectively elucidated the potential molecular mechanisms that collude symptoms, diseases, and molecules.

Data normalization for addressing the challenges in the analysis of single-cell transcriptomic datasets

BACKGROUND

Normalization is a critical step in the analysis of single-cell RNA-sequencing (scRNA-seq) datasets. Its main goal is to make gene counts comparable within and between cells. To do so, normalization methods must account for technical and biological variability. Numerous normalization methods have been developed addressing different sources of dispersion and making specific assumptions about the count data.

MAIN BODY

The selection of a normalization method has a direct impact on downstream analysis, for example differential gene expression and cluster identification. Thus, the objective of this review is to guide the reader in making an informed decision on the most appropriate normalization method to use. To this aim, we first give an overview of the different single cell sequencing platforms and methods commonly used including isolation and library preparation protocols. Next, we discuss the inherent sources of variability of scRNA-seq datasets. We describe the categories of normalization methods and include examples of each. We also delineate imputation and batch-effect correction methods. Furthermore, we describe data-driven metrics commonly used to evaluate the performance of normalization methods. We also discuss common scRNA-seq methods and toolkits used for integrated data analysis.

CONCLUSIONS

According to the correction performed, normalization methods can be broadly classified as within and between-sample algorithms. Moreover, with respect to the mathematical model used, normalization methods can further be classified into: global scaling methods, generalized linear models, mixed methods, and machine learning-based methods. Each of these methods depict pros and cons and make different statistical assumptions. However, there is no better performing normalization method. Instead, metrics such as silhouette width, K-nearest neighbor batch-effect test, or Highly Variable Genes are recommended to assess the performance of normalization methods.

scMGCN: A Multi-View Graph Convolutional Network for Cell Type Identification in scRNA-seq Data

Single-cell RNA sequencing (scRNA-seq) data reveal the complexity and diversity of cellular ecosystems and molecular interactions in various biomedical research. Hence, identifying cell types from large-scale scRNA-seq data using existing annotations is challenging and requires stable and interpretable methods. However, the current cell type identification methods have limited performance, mainly due to the intrinsic heterogeneity among cell populations and extrinsic differences between datasets. Here, we present a robust graph artificial intelligence model, a multi-view graph convolutional network model (scMGCN) that integrates multiple graph structures from raw scRNA-seq data and applies graph convolutional networks with attention mechanisms to learn cell embeddings and predict cell labels. We evaluate our model on single-dataset, cross-species, and cross-platform experiments and compare it with other state-of-the-art methods. Our results show that scMGCN outperforms the other methods regarding stability, accuracy, and robustness to batch effects. Our main contributions are as follows: Firstly, we introduce multi-view learning and multiple graph construction methods to capture comprehensive cellular information from scRNA-seq data. Secondly, we construct a scMGCN that combines graph convolutional networks with attention mechanisms to extract shared, high-order information from cells. Finally, we demonstrate the effectiveness and superiority of the scMGCN on various datasets.

Human pluripotent stem cell-derived inner ear organoids recapitulate otic development in vitro

Our molecular understanding of the early stages of human inner ear development has been limited by the difficulty in accessing fetal samples at early gestational stages. As an alternative, previous studies have shown that inner ear morphogenesis can be partially recapitulated using induced pluripotent stem cells directed to differentiate into inner ear organoids (IEOs). Once validated and benchmarked, these systems could represent unique tools to complement and refine our understanding of human otic differentiation and model developmental defects. Here, we provide the first direct comparisons of the early human embryonic otocyst and fetal sensory organs with human IEOs. We use multiplexed immunostaining and single-cell RNA-sequencing to characterize IEOs at three key developmental steps, providing a new and unique signature of in vitro-derived otic placode, epithelium, neuroblasts and sensory epithelia. In parallel, we evaluate the expression and localization of crucial markers at these equivalent stages in human embryos. Together, our data indicate that the current state-of-the-art protocol enables the specification of bona fide otic tissue, supporting the further application of IEOs to inform inner ear biology and disease.

Mapping oto-pharyngeal development in a human inner ear organoid model

Inner ear development requires the coordination of cell types from distinct epithelial, mesenchymal and neuronal lineages. Although we have learned much from animal models, many details about human inner ear development remain elusive. We recently developed an in vitro model of human inner ear organogenesis using pluripotent stem cells in a 3D culture, fostering the growth of a sensorineural circuit, including hair cells and neurons. Despite previously characterizing some cell types, many remain undefined. This study aimed to chart the in vitro development timeline of the inner ear organoid to understand the mechanisms at play. Using single-cell RNA sequencing at ten stages during the first 36 days of differentiation, we tracked the evolution from pluripotency to various ear cell types after exposure to specific signaling modulators. Our findings showcase gene expression that influences differentiation, identifying a plethora of ectodermal and mesenchymal cell types. We also discern aspects of the organoid model consistent with in vivo development, while highlighting potential discrepancies. Our study establishes the Inner Ear Organoid Developmental Atlas (IODA), offering deeper insights into human biology and improving inner ear tissue differentiation.

Generative modeling of single-cell gene expression for dose-dependent chemical perturbations

Single-cell sequencing reveals the heterogeneity of cellular response to chemical perturbations. However, testing all relevant combinations of cell types, chemicals, and doses is a daunting task. A deep generative learning formalism called variational autoencoders (VAEs) has been effective in predicting single-cell gene expression perturbations for single doses. Here, we introduce single-cell variational inference of dose-response (scVIDR), a VAE-based model that predicts both single-dose and multiple-dose cellular responses better than existing models. We show that scVIDR can predict dose-dependent gene expression across mouse hepatocytes, human blood cells, and cancer cell lines. We biologically interpret the latent space of scVIDR using a regression model and use scVIDR to order individual cells based on their sensitivity to chemical perturbation by assigning each cell a "pseudo-dose" value. We envision that scVIDR can help reduce the need for repeated animal testing across tissues, chemicals, and doses.

A single-cell level comparison of human inner ear organoids with the human cochlea and vestibular organs

Inner ear disorders are among the most common congenital abnormalities; however, current tissue culture models lack the cell type diversity to study these disorders and normal otic development. Here, we demonstrate the robustness of human pluripotent stem cell-derived inner ear organoids (IEOs) and evaluate cell type heterogeneity by single-cell transcriptomics. To validate our findings, we construct a single-cell atlas of human fetal and adult inner ear tissue. Our study identifies various cell types in the IEOs including periotic mesenchyme, type I and type II vestibular hair cells, and developing vestibular and cochlear epithelium. Many genes linked to congenital inner ear dysfunction are confirmed to be expressed in these cell types. Additional cell-cell communication analysis within IEOs and fetal tissue highlights the role of endothelial cells on the developing sensory epithelium. These findings provide insights into this organoid model and its potential applications in studying inner ear development and disorders.

Regionalized Protein Localization Domains in the Zebrafish Hair Cell Kinocilium

Sensory hair cells are the receptors for auditory, vestibular, and lateral line sensory organs in vertebrates. These cells are distinguished by "hair"-like projections from their apical surface collectively known as the hair bundle. Along with the staircase arrangement of the actin-filled stereocilia, the hair bundle features a single, non-motile, true cilium called the kinocilium. The kinocilium plays an important role in bundle development and the mechanics of sensory detection. To understand more about kinocilial development and structure, we performed a transcriptomic analysis of zebrafish hair cells to identify cilia-associated genes that have yet to be characterized in hair cells. In this study, we focused on three such genes-ankef1a, odf3l2a, and saxo2-because human or mouse orthologs are either associated with sensorineural hearing loss or are located near uncharacterized deafness loci. We made transgenic fish that express fluorescently tagged versions of their proteins, demonstrating their localization to the kinocilia of zebrafish hair cells. Furthermore, we found that Ankef1a, Odf3l2a, and Saxo2 exhibit distinct localization patterns along the length of the kinocilium and within the cell body. Lastly, we have reported a novel overexpression phenotype of Saxo2. Overall, these results suggest that the hair cell kinocilium in zebrafish is regionalized along its proximal-distal axis and set the groundwork to understand more about the roles of these kinocilial proteins in hair cells.

Human pluripotent stem cells-derived inner ear organoids recapitulate otic development in vitro

Our molecular understanding of the early stages of human inner ear development has been limited by the difficulty in accessing fetal samples at early gestational stages. As an alternative, previous studies have shown that inner ear morphogenesis can be partially recapitulated using induced pluripotent stem cells (iPSCs) directed to differentiate into Inner Ear Organoids (IEOs). Once validated and benchmarked, these systems could represent unique tools to complement and refine our understanding of human otic differentiation and model developmental defects. Here, we provide the first direct comparisons of the early human embryonic otocyst and human iPSC-derived IEOs. We use multiplexed immunostaining, and single-cell RNA sequencing to characterize IEOs at three key developmental steps, providing a new and unique signature of in vitro derived otic -placode, -epithelium, -neuroblasts, and -sensory epithelia. In parallel, we evaluate the expression and localization of critical markers at these equivalent stages in human embryos. We show that the placode derived in vitro (days 8-12) has similar marker expression to the developing otic placode of Carnegie Stage (CS) 11 embryos and subsequently (days 20-40) this gives rise to otic epithelia and neuroblasts comparable to the CS13 embryonic stage. Differentiation of sensory epithelia, including supporting cells and hair cells starts in vitro at days 50-60 of culture. The maturity of these cells is equivalent to vestibular sensory epithelia at week 10 or cochlear tissue at week 12 of development, before functional onset. Together, our data indicate that the current state-of-the-art protocol enables the specification of bona fide otic tissue, supporting the further application of IEOs to inform inner ear biology and disease.

The expression of p27 in the adult vestibular sensory organs and its possible roles

Vestibular hair cells (HCs) located in the inner ear are the receptors of vestibular sensory, which facilitates the human sense of balance. The detailed differentiation pattern and maturation process of the vestibular HCs are unclear now. p27, a cyclin/CDK inhibitor, plays a critical role in regulating the exit of cell cycle. We found that p27 was continuously expressed in the terminally differentiated and mature vestibular HCs using p27-P2A-iCreER/+; Rosa26-LSL-tdTomato/+ mice, suggesting p27 might have novel roles independent of its CDK inhibitory action. p27 is also reported to be associated with cell differentiation, cell migration and cell survival. We further explored the difference of p27 expression between two subtypes of vestibular HCs, and found that the proportion of p27-tdTomato positive type I vestibular HCs increased gradually along the subtype determination and maturation of vestibular HCs, suggesting that p27 might play a role in the HC subtype differentiation, maturation and function acquirement.

Adversarial dense graph convolutional networks for single-cell classification

MOTIVATION

In single-cell transcriptomics applications, effective identification of cell types in multicellular organisms and in-depth study of the relationships between genes has become one of the main goals of bioinformatics research. However, data heterogeneity and random noise pose significant difficulties for scRNA-seq data analysis.

RESULTS

We have proposed an adversarial dense graph convolutional network architecture for single-cell classification. Specifically, to enhance the representation of higher-order features and the organic combination between features, dense connectivity mechanism and attention-based feature aggregation are introduced for feature learning in convolutional neural networks. To preserve the features of the original data, we use a feature reconstruction module to assist the goal of single-cell classification. In addition, HNNVAT uses virtual adversarial training to improve the generalization and robustness. Experimental results show that our model outperforms the existing classical methods in terms of classification accuracy on benchmark datasets.

AVAILABILITY AND IMPLEMENTATION

The source code of HNNVAT is available at https://github.com/DisscLab/HNNVAT.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Single-cell transcriptomic profiling of the zebrafish inner ear reveals molecularly distinct hair cell and supporting cell subtypes

A major cause of human deafness and vestibular dysfunction is permanent loss of the mechanosensory hair cells of the inner ear. In non-mammalian vertebrates such as zebrafish, regeneration of missing hair cells can occur throughout life. While a comparative approach has the potential to reveal the basis of such differential regenerative ability, the degree to which the inner ears of fish and mammals share common hair cells and supporting cell types remains unresolved. Here, we perform single-cell RNA sequencing of the zebrafish inner ear at embryonic through adult stages to catalog the diversity of hair cells and non-sensory supporting cells. We identify a putative progenitor population for hair cells and supporting cells, as well as distinct hair and supporting cell types in the maculae versus cristae. The hair cell and supporting cell types differ from those described for the lateral line system, a distributed mechanosensory organ in zebrafish in which most studies of hair cell regeneration have been conducted. In the maculae, we identify two subtypes of hair cells that share gene expression with mammalian striolar or extrastriolar hair cells. In situ hybridization reveals that these hair cell subtypes occupy distinct spatial domains within the three macular organs, the utricle, saccule, and lagena, consistent with the reported distinct electrophysiological properties of hair cells within these domains. These findings suggest that primitive specialization of spatially distinct striolar and extrastriolar hair cells likely arose in the last common ancestor of fish and mammals. The similarities of inner ear cell type composition between fish and mammals validate zebrafish as a relevant model for understanding inner ear-specific hair cell function and regeneration.

Multiset multicover methods for discriminative marker selection

Markers are increasingly being used for several high-throughput data analysis and experimental design tasks. Examples include the use of markers for assigning cell types in scRNA-seq studies, for deconvolving bulk gene expression data, and for selecting marker proteins in single-cell spatial proteomics studies. Most marker selection methods focus on differential expression (DE) analysis. Although such methods work well for data with a few non-overlapping marker sets, they are not appropriate for large atlas-size datasets where several cell types and tissues are considered. To address this, we define the phenotype cover (PC) problem for marker selection and present algorithms that can improve the discriminative power of marker sets. Analysis of these sets on several marker-selection tasks suggests that these methods can lead to solutions that accurately distinguish different phenotypes in the data.

CHD7 regulates otic lineage specification and hair cell differentiation in human inner ear organoids

Mutations in CHD7 cause CHARGE syndrome, affecting multiple organs including the inner ear in humans. We investigate how CHD7 mutations affect inner ear development using human pluripotent stem cell-derived organoids as a model system. We find that loss of CHD7 or its chromatin remodeling activity leads to complete absence of hair cells and supporting cells, which can be explained by dysregulation of key otic development-associated genes in mutant otic progenitors. Further analysis of the mutant otic progenitors suggests that CHD7 can regulate otic genes through a chromatin remodeling-independent mechanism. Results from transcriptome profiling of hair cells reveal disruption of deafness gene expression as a potential underlying mechanism of CHARGE-associated sensorineural hearing loss. Notably, co-differentiating CHD7 knockout and wild-type cells in chimeric organoids partially rescues mutant phenotypes by restoring otherwise severely dysregulated otic genes. Taken together, our results suggest that CHD7 plays a critical role in regulating human otic lineage specification and hair cell differentiation.

The Genomics of Auditory Function and Disease

Current estimates suggest that nearly half a billion people worldwide are affected by hearing loss. Because of the major psychological, social, economic, and health ramifications, considerable efforts have been invested in identifying the genes and molecular pathways involved in hearing loss, whether genetic or environmental, to promote prevention, improve rehabilitation, and develop therapeutics. Genomic sequencing technologies have led to the discovery of genes associated with hearing loss. Studies of the transcriptome and epigenome of the inner ear have characterized key regulators and pathways involved in the development of the inner ear and have paved the way for their use in regenerative medicine. In parallel, the immense preclinical success of using viral vectors for gene delivery in animal models of hearing loss has motivated the industry to work on translating such approaches into the clinic. Here, we review the recent advances in the genomics of auditory function and dysfunction, from patient diagnostics to epigenetics and gene therapy.

Advancing discovery in hearing research via biologist-friendly access to multi-omic data

High-throughput cell type-specific multi-omic analyses have advanced our understanding of inner ear biology in an unprecedented way. The full benefit of these data, however, is reached from their re-use. Successful re-use of data requires identifying the natural users and ensuring proper data democratization and federation for their seamless and meaningful access. Here we discuss universal challenges in access and re-use of multi-omic data, possible solutions, and introduce the gEAR (the gene Expression Analysis Resource, umgear.org)-a tool for multi-omic data visualization, sharing and access for the ear field.

Cell-Type Resolved Insights into the Cis-Regulatory Genome of NAFLD

The prevalence of non-alcoholic fatty liver disease (NAFLD) is increasing rapidly, and unmet treatment can result in the development of hepatitis, fibrosis, and liver failure. There are difficulties involved in diagnosing NAFLD early and for this reason there are challenges involved in its treatment. Furthermore, no drugs are currently approved to alleviate complications, a fact which highlights the need for further insight into disease mechanisms. NAFLD pathogenesis is associated with complex cellular changes, including hepatocyte steatosis, immune cell infiltration, endothelial dysfunction, hepatic stellate cell activation, and epithelial ductular reaction. Many of these cellular changes are controlled by dramatic changes in gene expression orchestrated by the cis-regulatory genome and associated transcription factors. Thus, to understand disease mechanisms, we need extensive insights into the gene regulatory mechanisms associated with tissue remodeling. Mapping cis-regulatory regions genome-wide is a step towards this objective and several current and emerging technologies allow detection of accessible chromatin and specific histone modifications in enriched cell populations of the liver, as well as in single cells. Here, we discuss recent insights into the cis-regulatory genome in NAFLD both at the organ-level and in specific cell populations of the liver. Moreover, we highlight emerging technologies that enable single-cell resolved analysis of the cis-regulatory genome of the liver.