RNA Sequencing of Single Human Islet Cells Reveals Type 2 Diabetes Genes

scQA: A dual-perspective cell type identification model for single cell transcriptome data

Single-cell RNA sequencing technologies have been pivotal in advancing the development of algorithms for clustering heterogeneous cell populations. Existing methods for utilizing scRNA-seq data to identify cell types tend to neglect the beneficial impact of dropout events and perform clustering focusing solely on quantitative perspective. Here, we introduce a novel method named scQA, notable for its ability to concurrently identify cell types and cell type-specific key genes from both qualitative and quantitative perspectives. In contrast to other methods, scQA not only identifies cell types but also extracts key genes associated with these cell types, enabling bidirectional clustering for scRNA-seq data. Through an iterative process, our approach aims to minimize the number of landmarks to approximately a dozen while maximizing the inclusion of quasi-trend-preserved genes with dropouts both qualitatively and quantitatively. It then clusters cells by employing an ingenious label propagation strategy, obviating the requirement for a predetermined number of cell types. Validated on 20 publicly available scRNA-seq datasets, scQA consistently outperforms other salient tools. Furthermore, we confirm the effectiveness and potential biological significance of the identified key genes through both external and internal validation. In conclusion, scQA emerges as a valuable tool for investigating cell heterogeneity due to its distinctive fusion of qualitative and quantitative facets, along with bidirectional clustering capabilities. Furthermore, it can be seamlessly integrated into border scRNA-seq analyses. The source codes are publicly available at https://github.com/LD-Lyndee/scQA.

Mammalian START-like phosphatidylinositol transfer proteins - Physiological perspectives and roles in cancer biology

PtdIns and its phosphorylated derivatives, the phosphoinositides, are the biochemical components of a major pathway of intracellular signaling in all eukaryotic cells. These lipids are few in terms of cohort of unique positional isomers, and are quantitatively minor species of the bulk cellular lipidome. Nevertheless, phosphoinositides regulate an impressively diverse set of biological processes. It is from that perspective that perturbations in phosphoinositide-dependent signaling pathways are increasingly being recognized as causal foundations of many human diseases - including cancer. Although phosphatidylinositol transfer proteins (PITPs) are not enzymes, these proteins are physiologically significant regulators of phosphoinositide signaling. As such, PITPs are conserved throughout the eukaryotic kingdom. Their biological importance notwithstanding, PITPs remain understudied. Herein, we review current information regarding PITP biology primarily focusing on how derangements in PITP function disrupt key signaling/developmental pathways and are associated with a growing list of pathologies in mammals.

Single cell glucose-stimulated insulin secretion assay using nanowell-in-microwell plates

Pancreatic β cells secrete insulin in response to elevated levels of glucose. Stem cell derived β (SCβ) cells aim to replicate this glucose-stimulated insulin secretion (GSIS) function, but current preparations cannot provide the same level of insulin as natural β cells. Here, we develop an assay to measure GSIS at the single cell level to investigate the functional heterogeneity of SCβ cells and donor-derived islet cells. Our assay involves randomly depositing single cells and insulin capture microbeads in open-top nanowells (40 × 40 × 55 μm3) fabricated on glass-bottom imaging microwell plates. Insulin secreted from single cells is captured on microbeads and then stained using a detection antibody. The nanowell microstructure limits diffusion of secreted insulin. The glass substrate provides an optically flat surface for quantitative microscopy to measure the concentration of secreted insulin. We used this approach to measure GSIS from SCβ cells and donor-derived islet cells after 15 minutes exposure to 3.3 mM and 16.7 mM glucose. Both cell types exhibited significant GSIS heterogeneity, where elite cells (<20%) produced the majority of the secreted insulin (55-78%). This assay provides an immediate readout of single cell glucose-stimulated insulin secretion in a flexible well plate-based format.

XgCPred: Cell type classification using XGBoost-CNN integration and exploiting gene expression imaging in single-cell RNAseq data

BACKGROUND

The advent of single-cell RNA sequencing (scRNA-seq) has revolutionized transcriptomic research by enabling the exploration of gene expression at an individual cell level. This advancement sheds light on how cells differentiate and evolve over time. Effectively classification cell types within scRNA-seq datasets are essential for understanding the intricate cell compositions within tissues and elucidating the origins of various diseases. Challenges persist in the field, emphasizing the need for precise categorization across diverse datasets, addressing aggregated cell data, and managing the complexity of high-dimensional data spaces.

METHODOLOGY

XgCPred is a novel approach combining XGBoost with convolutional neural networks (CNNs) to provide cell type classification with better accuracy in single-cell RNA-seq data. This combo reveals how well CNNs can detect spatial hierarchy in gene expression images and how XGBoost performs with large volumes of data. XgCPred utilizes an imaging representation of gene expression that is based on the hierarchical organization of genes found in the KEGG BRITE database.

RESULTS

Rigorous testing of XgCPred across multiple scRNA-seq datasets, each presenting unique challenges such as varying cell counts, gene expression diversity, and cellular heterogeneity, has demonstrated its superiority compared to earlier methods. The algorithm shows remarkable accuracy and precision in cell type annotation, achieving near-perfect classification scores in some cases. These results underscore its capability to effectively manage data variability.

CONCLUSIONS

XgCPred distinguishes itself through its dependable and accurate cell type classification across a range of scRNA-seq datasets. Its effectiveness stems from sophisticated data handling and its ability to adapt to the complexities inherent in scRNA-seq data. XgCPred delivers reliable cell annotations essential for further biological analysis and research, marking a significant advancement in genomic studies. With scRNA-seq datasets growing in size and complexity, XgCPred offers a scalable and potent solution for cell type identification, potentially enhancing our understanding of cellular biology and aiding in the precise detection of diseases. XgCPred is a useful tool in genomic research and tailored therapy because it solves current constraints on computing efficiency and generalizability.

COMSE: analysis of single-cell RNA-seq data using community detection-based feature selection

BACKGROUND

Single-cell RNA sequencing enables studying cells individually, yet high gene dimensions and low cell numbers challenge analysis. And only a subset of the genes detected are involved in the biological processes underlying cell-type specific functions.

RESULT

In this study, we present COMSE, an unsupervised feature selection framework using community detection to capture informative genes from scRNA-seq data. COMSE identified homogenous cell substates with high resolution, as demonstrated by distinguishing different cell cycle stages. Evaluations based on real and simulated scRNA-seq datasets showed COMSE outperformed methods even with high dropout rates in cell clustering assignment. We also demonstrate that by identifying communities of genes associated with batch effects, COMSE parses signals reflecting biological difference from noise arising due to differences in sequencing protocols, thereby enabling integrated analysis of scRNA-seq datasets of different sources.

CONCLUSIONS

COMSE provides an efficient unsupervised framework that selects highly informative genes in scRNA-seq data improving cell sub-states identification and cell clustering. It identifies gene subsets that reveal biological and technical heterogeneity, supporting applications like batch effect correction and pathway analysis. It also provides robust results for bulk RNA-seq data analysis.

scPriorGraph: constructing biosemantic cell-cell graphs with prior gene set selection for cell type identification from scRNA-seq data

Cell type identification is an indispensable analytical step in single-cell data analyses. To address the high noise stemming from gene expression data, existing computational methods often overlook the biologically meaningful relationships between genes, opting to reduce all genes to a unified data space. We assume that such relationships can aid in characterizing cell type features and improving cell type recognition accuracy. To this end, we introduce scPriorGraph, a dual-channel graph neural network that integrates multi-level gene biosemantics. Experimental results demonstrate that scPriorGraph effectively aggregates feature values of similar cells using high-quality graphs, achieving state-of-the-art performance in cell type identification.

Assessing parameter efficient methods for pre-trained language model in annotating scRNA-seq data

Annotating cell types of single-cell RNA sequencing (scRNA-seq) data is crucial for studying cellular heterogeneity in the tumor microenvironment. Recently, large-scale pre-trained language models (PLMs) have achieved significant progress in cell-type annotation of scRNA-seq data. This approach effectively addresses previous methods' shortcomings in performance and generalization. However, fine-tuning PLMs for different downstream tasks demands considerable computational resources, rendering it impractical. Hence, a new research branch introduces parameter-efficient fine-tuning (PEFT). This involves optimizing a few parameters while leaving the majority unchanged, leading to substantial reductions in computational expenses. Here, we utilize scBERT, a large-scale pre-trained model, to explore the capabilities of three PEFT methods in scRNA-seq cell type annotation. Extensive benchmark studies across several datasets demonstrate the superior applicability of PEFT methods. Furthermore, downstream analysis using models obtained through PEFT showcases their utility in novel cell type discovery and model interpretability for potential marker genes. Our findings underscore the considerable potential of PEFT in PLM-based cell type annotation, presenting novel perspectives for the analysis of scRNA-seq data.

scHD4E: Novel ensemble learning-based differential expression analysis method for single-cell RNA-sequencing data

Differential expression (DE) analysis between cell types for scRNA-seq data by capturing its complicated features is crucial. Recently, different methods have been developed for targeting the scRNA-seq data analysis based on different modeling frameworks, assumptions, strategies and test statistic in considering various data features. The scDEA is an ensemble learning-based DE analysis method developed recently, yielding p-values using Lancaster's combination, generated by 12 individual DE analysis methods, and producing more accurate and stable results than individual methods. The objective of our study is to propose a new ensemble learning-based DE analysis method, scHD4E, using top performers in only 4 separate methods. The top performer 4 methods have been selected through an evaluation process using six real scRNA-seq data sets. We conducted comprehensive experiments for five experimental data sets to evaluate our proposed method based on the sample size effects, batch effects, type I error control, gene ontology enrichment analysis, runtime, identified matched DE genes, and semantic similarity measurement between methods. We also perform similar analyses (except the last 3 terms) and compute performance measures like accuracy, F1 score, Mathew's correlation coefficient etc. for a simulated data set. The results show that scHD4E is performs better than all the individual and scDEA methods in all the above perspectives. We expect that scHD4E will serve the modern data scientists for detecting the DEGs in scRNA-seq data analysis. To implement our proposed method, a Github R package scHD4E and its shiny application has been developed, and available in the following links: https://github.com/bbiswas1989/scHD4E and https://github.com/bbiswas1989/scHD4E-Shiny.

Single-Cell RNA-seq Analysis Reveals a Positive Correlation between Ferroptosis and Beta-Cell Dedifferentiation in Type 2 Diabetes

Insulin deficiency in patients with type 2 diabetes mellitus (T2D) is associated with beta-cell dysfunction, a condition increasingly recognized to involve processes such as dedifferentiation and apoptosis. Moreover, emerging research points to a potential role for ferroptosis in the pathogenesis of T2D. In this study, we aimed to investigate the potential involvement of ferroptosis in the dedifferentiation of beta cells in T2D. We performed single-cell RNA sequencing analysis of six public datasets. Differential expression and gene set enrichment analyses were carried out to investigate the role of ferroptosis. Gene set variation and pseudo-time trajectory analyses were subsequently used to verify ferroptosis-related beta clusters. After cells were categorized according to their ferroptosis and dedifferentiation scores, we constructed transcriptional and competitive endogenous RNA networks, and validated the hub genes via machine learning and immunohistochemistry. We found that ferroptosis was enriched in T2D beta cells and that there was a positive correlation between ferroptosis and the process of dedifferentiation. Upon further analysis, we identified two beta clusters that presented pronounced features associated with ferroptosis and dedifferentiation. Several key transcription factors and 2 long noncoding RNAs (MALAT1 and MEG3) were identified. Finally, we confirmed that ferroptosis occurred in the pancreas of high-fat diet-fed mice and identified 4 proteins (NFE2L2, CHMP5, PTEN, and STAT3) that may participate in the effect of ferroptosis on dedifferentiation. This study helps to elucidate the interplay between ferroptosis and beta-cell health and opens new avenues for developing therapeutic strategies to treat diabetes.

Integrating Mendelian randomization and single-cell RNA sequencing to identify therapeutic targets of baicalin for type 2 diabetes mellitus

Background

Alternative and complementary therapies play an imperative role in the clinical management of Type 2 diabetes mellitus (T2DM), and exploring and utilizing natural products from a genetic perspective may yield novel insights into the mechanisms and interventions of the disorder.

Methods

To identify the therapeutic target of baicalin for T2DM, we conducted a Mendelian randomization study. Druggable targets of baicalin were obtained by integrating multiple databases, and target-associated cis-expression quantitative trait loci (cis-eQTL) originated from the eQTLGen consortium. Summary statistics for T2DM were derived from two independent genome-wide association studies available through the DIAGRAM Consortium (74,124 cases vs. 824,006 controls) and the FinnGen R9 repository (9,978 cases vs. 12,348 controls). Network construction and enrichment analysis were applied to the therapeutic targets of baicalin. Colocalization analysis was utilized to assess the potential for the therapeutic targets and T2DM to share causative genetic variations. Molecular docking was performed to validate the potency of baicalin. Single-cell RNA sequencing was employed to seek evidence of therapeutic targets' involvement in islet function.

Results

Eight baicalin-related targets proved to be significant in the discovery and validation cohorts. Genetic evidence indicated the expression of ANPEP, BECN1, HNF1A, and ST6GAL1 increased the risk of T2DM, and the expression of PGF, RXRA, SREBF1, and USP7 decreased the risk of T2DM. In particular, SREBF1 has significant interaction properties with other therapeutic targets and is supported by strong colocalization. Baicalin had favorable combination activity with eight therapeutic targets. The expression patterns of the therapeutic targets were characterized in cellular clusters of pancreatic tissues that exhibited a pseudo-temporal dependence on islet cell formation and development.

Conclusion

This study identified eight potential targets of baicalin for treating T2DM from a genetic perspective, contributing an innovative analytical framework for the development of natural products. We have offered fresh insights into the connections between therapeutic targets and islet cells. Further, fundamental experiments and clinical research are warranted to delve deeper into the molecular mechanisms of T2DM.

Detecting novel cell type in single-cell chromatin accessibility data via open-set domain adaptation

Recent advances in single-cell technologies enable the rapid growth of multi-omics data. Cell type annotation is one common task in analyzing single-cell data. It is a challenge that some cell types in the testing set are not present in the training set (i.e. unknown cell types). Most scATAC-seq cell type annotation methods generally assign each cell in the testing set to one known type in the training set but neglect unknown cell types. Here, we present OVAAnno, an automatic cell types annotation method which utilizes open-set domain adaptation to detect unknown cell types in scATAC-seq data. Comprehensive experiments show that OVAAnno successfully identifies known and unknown cell types. Further experiments demonstrate that OVAAnno also performs well on scRNA-seq data. Our codes are available online at https://github.com/lisaber/OVAAnno/tree/master.

A comparison of scRNA-seq annotation methods based on experimentally labeled immune cell subtype dataset

Cell-type annotation is a critical step in single-cell data analysis. With the development of numerous cell annotation methods, it is necessary to evaluate these methods to help researchers use them effectively. Reference datasets are essential for evaluation, but currently, the cell labels of reference datasets mainly come from computational methods, which may have computational biases and may not reflect the actual cell-type outcomes. This study first constructed an experimentally labeled immune cell-subtype single-cell dataset of the same batch and systematically evaluated 18 cell annotation methods. We assessed those methods under five scenarios, including intra-dataset validation, immune cell-subtype validation, unsupervised clustering, inter-dataset annotation, and unknown cell-type prediction. Accuracy and ARI were evaluation metrics. The results showed that SVM, scBERT, and scDeepSort were the best-performing supervised methods. Seurat was the best-performing unsupervised clustering method, but it couldn't fully fit the actual cell-type distribution. Our results indicated that experimentally labeled immune cell-subtype datasets revealed the deficiencies of unsupervised clustering methods and provided new dataset support for supervised methods.

Beaconet: A Reference-Free Method for Integrating Multiple Batches of Single-Cell Transcriptomic Data in Original Molecular Space

Integrating multiple single-cell datasets is essential for the comprehensive understanding of cell heterogeneity. Batch effect is the undesired systematic variations among technologies or experimental laboratories that distort biological signals and hinder the integration of single-cell datasets. However, existing methods typically rely on a selected dataset as a reference, leading to inconsistent integration performance using different references, or embed cells into uninterpretable low-dimensional feature space. To overcome these limitations, a reference-free method, Beaconet, for integrating multiple single-cell transcriptomic datasets in original molecular space by aligning the global distribution of each batch using an adversarial correction network is presented. Through extensive comparisons with 13 state-of-the-art methods, it is demonstrated that Beaconet can effectively remove batch effect while preserving biological variations and is superior to existing unsupervised methods using all possible references in overall performance. Furthermore, Beaconet performs integration in the original molecular feature space, enabling the characterization of cell types and downstream differential expression analysis directly using integrated data with gene-expression features. Additionally, when applying to large-scale atlas data integration, Beaconet shows notable advantages in both time- and space-efficiencies. In summary, Beaconet serves as an effective and efficient batch effect removal tool that can facilitate the integration of single-cell datasets in a reference-free and molecular feature-preserved mode.

Biochemical and metabolic characterization of a G6PC2 inhibitor

Three glucose-6-phosphatase catalytic subunits, that hydrolyze glucose-6-phosphate (G6P) to glucose and inorganic phosphate, have been identified, designated G6PC1-3, but only G6PC1 and G6PC2 have been implicated in the regulation of fasting blood glucose (FBG). Elevated FBG has been associated with multiple adverse clinical outcomes, including increased risk for type 2 diabetes and various cancers. Therefore, G6PC1 and G6PC2 inhibitors that lower FBG may be of prophylactic value for the prevention of multiple conditions. The studies described here characterize a G6PC2 inhibitor, designated VU0945627, previously identified as Compound 3. We show that VU0945627 preferentially inhibits human G6PC2 versus human G6PC1 but activates human G6PC3. VU0945627 is a mixed G6PC2 inhibitor, increasing the Km but reducing the Vmax for G6P hydrolysis. PyRx virtual docking to an AlphaFold2-derived G6PC2 structural model suggests VU0945627 binds two sites in human G6PC2. Mutation of residues in these sites reduces the inhibitory effect of VU0945627. VU0945627 does not inhibit mouse G6PC2 despite its 84% sequence identity with human G6PC2. Mutagenesis studies suggest this lack of inhibition of mouse G6PC2 is due, in part, to a change in residue 318 from histidine in human G6PC2 to proline in mouse G6PC2. Surprisingly, VU0945627 still inhibited glucose cycling in the mouse islet-derived βTC-3 cell line. Studies using intact mouse liver microsomes and PyRx docking suggest that this observation can be explained by an ability of VU0945627 to also inhibit the G6P transporter SLC37A4. These data will inform future computational modeling studies designed to identify G6PC isoform-specific inhibitors.

PredGCN: a Pruning-enabled Gene-Cell Net for automatic cell annotation of single cell transcriptome data

MOTIVATION

The annotation of cell types from single-cell transcriptomics is essential for understanding the biological identity and functionality of cellular populations. Although manual annotation remains the gold standard, the advent of automatic pipelines has become crucial for scalable, unbiased, and cost-effective annotations. Nonetheless, the effectiveness of these automatic methods, particularly those employing deep learning, significantly depends on the architecture of the classifier and the quality and diversity of the training datasets.

RESULTS

To address these limitations, we present a Pruning-enabled Gene-Cell Net (PredGCN) incorporating a Coupled Gene-Cell Net (CGCN) to enable representation learning and information storage. PredGCN integrates a Gene Splicing Net (GSN) and a Cell Stratification Net (CSN), employing a pruning operation (PrO) to dynamically tackle the complexity of heterogeneous cell identification. Among them, GSN leverages multiple statistical and hypothesis-driven feature extraction methods to selectively assemble genes with specificity for scRNA-seq data while CSN unifies elements based on diverse region demarcation principles, exploiting the representations from GSN and precise identification from different regional homogeneity perspectives. Furthermore, we develop a multi-objective Pareto pruning operation (Pareto PrO) to expand the dynamic capabilities of CGCN, optimizing the sub-network structure for accurate cell type annotation. Multiple comparison experiments on real scRNA-seq datasets from various species have demonstrated that PredGCN surpasses existing state-of-the-art methods, including its scalability to cross-species datasets. Moreover, PredGCN can uncover unknown cell types and provide functional genomic analysis by quantifying the influence of genes on cell clusters, bringing new insights into cell type identification and characterizing scRNA-seq data from different perspectives.

AVAILABILITY AND IMPLEMENTATION

The source code is available at https://github.com/IrisQi7/PredGCN and test data is available at https://figshare.com/articles/dataset/PredGCN/25251163.

Optogenetic β cell interrogation in vivo reveals a functional hierarchy directing the Ca2+ response to glucose supported by vitamin B6

Coordination of cellular activity through Ca2+ enables β cells to secrete precise quantities of insulin. To explore how the Ca2+ response is orchestrated in space and time, we implement optogenetic systems to probe the role of individual β cells in the glucose response. By targeted β cell activation/inactivation in zebrafish, we reveal a hierarchy of cells, each with a different level of influence over islet-wide Ca2+ dynamics. First-responder β cells lie at the top of the hierarchy, essential for initiating the first-phase Ca2+ response. Silencing first responders impairs the Ca2+ response to glucose. Conversely, selective activation of first responders demonstrates their increased capability to raise pan-islet Ca2+ levels compared to followers. By photolabeling and transcriptionally profiling β cells that differ in their thresholds to a glucose-stimulated Ca2+ response, we highlight vitamin B6 production as a signature pathway of first responders. We further define an evolutionarily conserved requirement for vitamin B6 in enabling the Ca2+ response to glucose in mammalian systems.

Regulation of endocrine cell alternative splicing revealed by single-cell RNA sequencing in type 2 diabetes pathogenesis

The prevalent RNA alternative splicing (AS) contributes to molecular diversity, which has been demonstrated in cellular function regulation and disease pathogenesis. However, the contribution of AS in pancreatic islets during diabetes progression remains unclear. Here, we reanalyze the full-length single-cell RNA sequencing data from the deposited database to investigate AS regulation across human pancreatic endocrine cell types in non-diabetic (ND) and type 2 diabetic (T2D) individuals. Our analysis demonstrates the significant association between transcriptomic AS profiles and cell-type-specificity, which could be applied to distinguish the clustering of major endocrine cell types. Moreover, AS profiles are enabled to clearly define the mature subset of β-cells in healthy controls, which is completely lost in T2D. Further analysis reveals that RNA-binding proteins (RBPs), heterogeneous nuclear ribonucleoproteins (hnRNPs) and FXR1 family proteins are predicted to induce the functional impairment of β-cells through regulating AS profiles. Finally, trajectory analysis of endocrine cells suggests the β-cell identity shift through dedifferentiation and transdifferentiation of β-cells during the progression of T2D. Together, our study provides a mechanism for regulating β-cell functions and suggests the significant contribution of AS program during diabetes pathogenesis.

Predicting type 2 diabetes via machine learning integration of multiple omics from human pancreatic islets

Type 2 diabetes (T2D) is the fastest growing non-infectious disease worldwide. Impaired insulin secretion from pancreatic beta-cells is a hallmark of T2D, but the mechanisms behind this defect are insufficiently characterized. Integrating multiple layers of biomedical information, such as different Omics, may allow more accurate understanding of complex diseases such as T2D. Our aim was to explore and use Machine Learning to integrate multiple sources of biological/molecular information (multiOmics), in our case RNA-sequening, DNA methylation, SNP and phenotypic data from islet donors with T2D and non-diabetic controls. We exploited Machine Learning to perform multiOmics integration of DNA methylation, expression, SNPs, and phenotypes from pancreatic islets of 110 individuals, with ~ 30% being T2D cases. DNA methylation was analyzed using Infinium MethylationEPIC array, expression was analyzed using RNA-sequencing, and SNPs were analyzed using HumanOmniExpress arrays. Supervised linear multiOmics integration via DIABLO based on Partial Least Squares (PLS) achieved an accuracy of 91 ± 15% of T2D prediction with an area under the curve of 0.96 ± 0.08 on the test dataset after cross-validation. Biomarkers identified by this multiOmics integration, including SACS and TXNIP DNA methylation, OPRD1 and RHOT1 expression and a SNP annotated to ANO1, provide novel insights into the interplay between different biological mechanisms contributing to T2D. This Machine Learning approach of multiOmics cross-sectional data from human pancreatic islets achieved a promising accuracy of T2D prediction, which may potentially find broad applications in clinical diagnostics. In addition, it delivered novel candidate biomarkers for T2D and links between them across the different Omics.

HumanIslets: An integrated platform for human islet data access and analysis

Comprehensive molecular and cellular phenotyping of human islets can enable deep mechanistic insights for diabetes research. We established the Human Islet Data Analysis and Sharing (HI-DAS) consortium to advance goals in accessibility, usability, and integration of data from human islets isolated from donors with and without diabetes at the Alberta Diabetes Institute (ADI) IsletCore. Here we introduce HumanIslets.com, an open resource for the research community. This platform, which presently includes data on 547 human islet donors, allows users to access linked datasets describing molecular profiles, islet function and donor phenotypes, and to perform various statistical and functional analyses at the donor, islet and single-cell levels. As an example of the analytic capacity of this resource we show a dissociation between cell culture effects on transcript and protein expression, and an approach to correct for exocrine contamination found in hand-picked islets. Finally, we provide an example workflow and visualization that highlights links between type 2 diabetes status, SERCA3b Ca2+-ATPase levels at the transcript and protein level, insulin secretion and islet cell phenotypes. HumanIslets.com provides a growing and adaptable set of resources and tools to support the metabolism and diabetes research community.

SlowMoMan: a web app for discovery of important features along user-drawn trajectories in 2D embeddings

Motivation

Nonlinear low-dimensional embeddings allow humans to visualize high-dimensional data, as is often seen in bioinformatics, where datasets may have tens of thousands of dimensions. However, relating the axes of a nonlinear embedding to the original dimensions is a nontrivial problem. In particular, humans may identify patterns or interesting subsections in the embedding, but cannot easily identify what those patterns correspond to in the original data.

Results

Thus, we present SlowMoMan (SLOW Motions on MANifolds), a web application which allows the user to draw a one-dimensional path onto a 2D embedding. Then, by back-projecting the manifold to the original, high-dimensional space, we sort the original features such that those most discriminative along the manifold are ranked highly. We show a number of pertinent use cases for our tool, including trajectory inference, spatial transcriptomics, and automatic cell classification.

Availability and implementation

Software: https://yunwilliamyu.github.io/SlowMoMan/; Code: https://github.com/yunwilliamyu/SlowMoMan.

Endoplasmic reticulum stress in pancreatic β-cell dysfunctionality and diabetes mellitus: a promising target for generation of functional hPSC-derived β-cells in vitro

Diabetes mellitus (DM), is a chronic disorder characterized by impaired glucose homeostasis that results from the loss or dysfunction of pancreatic β-cells leading to type 1 diabetes (T1DM) and type 2 diabetes (T2DM), respectively. Pancreatic β-cells rely to a great degree on their endoplasmic reticulum (ER) to overcome the increased secretary need for insulin biosynthesis and secretion in response to nutrient demand to maintain glucose homeostasis in the body. As a result, β-cells are potentially under ER stress following nutrient levels rise in the circulation for a proper pro-insulin folding mediated by the unfolded protein response (UPR), underscoring the importance of this process to maintain ER homeostasis for normal β-cell function. However, excessive or prolonged increased influx of nascent proinsulin into the ER lumen can exceed the ER capacity leading to pancreatic β-cells ER stress and subsequently to β-cell dysfunction. In mammalian cells, such as β-cells, the ER stress response is primarily regulated by three canonical ER-resident transmembrane proteins: ATF6, IRE1, and PERK/PEK. Each of these proteins generates a transcription factor (ATF4, XBP1s, and ATF6, respectively), which in turn activates the transcription of ER stress-inducible genes. An increasing number of evidence suggests that unresolved or dysregulated ER stress signaling pathways play a pivotal role in β-cell failure leading to insulin secretion defect and diabetes. In this article we first highlight and summarize recent insights on the role of ER stress and its associated signaling mechanisms on β-cell function and diabetes and second how the ER stress pathways could be targeted in vitro during direct differentiation protocols for generation of hPSC-derived pancreatic β-cells to faithfully phenocopy all features of bona fide human β-cells for diabetes therapy or drug screening.

scPLAN: a hierarchical computational framework for single transcriptomics data annotation, integration and cell-type label refinement

MOTIVATION

In the past decade, single-cell RNA sequencing (scRNA-seq) has emerged as a pivotal method for transcriptomic profiling in biomedical research. Precise cell-type identification is crucial for subsequent analysis of single-cell data. And the integration and refinement of annotated data are essential for building comprehensive databases. However, prevailing annotation techniques often overlook the hierarchical organization of cell types, resulting in inconsistent annotations. Meanwhile, most existing integration approaches fail to integrate datasets with different annotation depths and none of them can enhance the labels of outdated data with lower annotation resolutions using more intricately annotated datasets or novel biological findings.

RESULTS

Here, we introduce scPLAN, a hierarchical computational framework designed for scRNA-seq data analysis. scPLAN excels in annotating unlabeled scRNA-seq data using a reference dataset structured along a hierarchical cell-type tree. It identifies potential novel cell types in a systematic, layer-by-layer manner. Additionally, scPLAN effectively integrates annotated scRNA-seq datasets with varying levels of annotation depth, ensuring consistent refinement of cell-type labels across datasets with lower resolutions. Through extensive annotation and novel cell detection experiments, scPLAN has demonstrated its efficacy. Two case studies have been conducted to showcase how scPLAN integrates datasets with diverse cell-type label resolutions and refine their cell-type labels.

AVAILABILITY

https://github.com/michaelGuo1204/scPLAN.

Comprehensive single-cell RNA-seq analysis using deep interpretable generative modeling guided by biological hierarchy knowledge

Recent advances in microfluidics and sequencing technologies allow researchers to explore cellular heterogeneity at single-cell resolution. In recent years, deep learning frameworks, such as generative models, have brought great changes to the analysis of transcriptomic data. Nevertheless, relying on the potential space of these generative models alone is insufficient to generate biological explanations. In addition, most of the previous work based on generative models is limited to shallow neural networks with one to three layers of latent variables, which may limit the capabilities of the models. Here, we propose a deep interpretable generative model called d-scIGM for single-cell data analysis. d-scIGM combines sawtooth connectivity techniques and residual networks, thereby constructing a deep generative framework. In addition, d-scIGM incorporates hierarchical prior knowledge of biological domains to enhance the interpretability of the model. We show that d-scIGM achieves excellent performance in a variety of fundamental tasks, including clustering, visualization, and pseudo-temporal inference. Through topic pathway studies, we found that d-scIGM-learned topics are better enriched for biologically meaningful pathways compared to the baseline models. Furthermore, the analysis of drug response data shows that d-scIGM can capture drug response patterns in large-scale experiments, which provides a promising way to elucidate the underlying biological mechanisms. Lastly, in the melanoma dataset, d-scIGM accurately identified different cell types and revealed multiple melanin-related driver genes and key pathways, which are critical for understanding disease mechanisms and drug development.

NKX2-2 based nuclei sorting on frozen human archival pancreas enables the enrichment of islet endocrine populations for single-nucleus RNA sequencing

BACKGROUND

Current approaches to profile the single-cell transcriptomics of human pancreatic endocrine cells almost exclusively rely on freshly isolated islets. However, human islets are limited in availability. Furthermore, the extensive processing steps during islet isolation and subsequent single cell dissolution might alter gene expressions. In this work, we report the development of a single-nucleus RNA sequencing (snRNA-seq) approach with targeted islet cell enrichment for endocrine-population focused transcriptomic profiling using frozen archival pancreatic tissues without islet isolation.

RESULTS

We cross-compared five nuclei isolation protocols and selected the citric acid method as the best strategy to isolate nuclei with high RNA integrity and low cytoplasmic contamination from frozen archival human pancreata. We innovated fluorescence-activated nuclei sorting based on the positive signal of NKX2-2 antibody to enrich nuclei of the endocrine population from the entire nuclei pool of the pancreas. Our sample preparation procedure generated high-quality single-nucleus gene-expression libraries while preserving the endocrine population diversity. In comparison with single-cell RNA sequencing (scRNA-seq) library generated with live cells from freshly isolated human islets, the snRNA-seq library displayed comparable endocrine cellular composition and cell type signature gene expression. However, between these two types of libraries, differential enrichments of transcripts belonging to different functional classes could be observed.

CONCLUSIONS

Our work fills a technological gap and helps to unleash frozen archival pancreatic tissues for molecular profiling targeting the endocrine population. This study opens doors to retrospective mappings of endocrine cell dynamics in pancreatic tissues of complex histopathology. We expect that our protocol is applicable to enrich nuclei for transcriptomics studies from various populations in different types of frozen archival tissues.

DISSECT: deep semi-supervised consistency regularization for accurate cell type fraction and gene expression estimation

Cell deconvolution is the estimation of cell type fractions and cell type-specific gene expression from mixed data. An unmet challenge in cell deconvolution is the scarcity of realistic training data and the domain shift often observed in synthetic training data. Here, we show that two novel deep neural networks with simultaneous consistency regularization of the target and training domains significantly improve deconvolution performance. Our algorithm, DISSECT, outperforms competing algorithms in cell fraction and gene expression estimation by up to 14 percentage points. DISSECT can be easily adapted to other biomedical data types, as exemplified by our proteomic deconvolution experiments.

Pancreatic β-Cell Identity Change through the Lens of Single-Cell Omics Research

The main hallmark in the development of both type 1 and type 2 diabetes is a decline in functional β-cell mass. This decline is predominantly attributed to β-cell death, although recent findings suggest that the loss of β-cell identity may also contribute to β-cell dysfunction. This phenomenon is characterized by a reduced expression of key markers associated with β-cell identity. This review delves into the insights gained from single-cell omics research specifically focused on β-cell identity. It highlights how single-cell omics based studies have uncovered an unexpected level of heterogeneity among β-cells and have facilitated the identification of distinct β-cell subpopulations through the discovery of cell surface markers, transcriptional regulators, the upregulation of stress-related genes, and alterations in chromatin activity. Furthermore, specific subsets of β-cells have been identified in diabetes, such as displaying an immature, dedifferentiated gene signature, expressing significantly lower insulin mRNA levels, and expressing increased β-cell precursor markers. Additionally, single-cell omics has increased insight into the detrimental effects of diabetes-associated conditions, including endoplasmic reticulum stress, oxidative stress, and inflammation, on β-cell identity. Lastly, this review outlines the factors that may influence the identification of β-cell subpopulations when designing and performing a single-cell omics experiment.

LDHB contributes to the regulation of lactate levels and basal insulin secretion in human pancreatic β cells

Using 13C6 glucose labeling coupled to gas chromatography-mass spectrometry and 2D 1H-13C heteronuclear single quantum coherence NMR spectroscopy, we have obtained a comparative high-resolution map of glucose fate underpinning β cell function. In both mouse and human islets, the contribution of glucose to the tricarboxylic acid (TCA) cycle is similar. Pyruvate fueling of the TCA cycle is primarily mediated by the activity of pyruvate dehydrogenase, with lower flux through pyruvate carboxylase. While the conversion of pyruvate to lactate by lactate dehydrogenase (LDH) can be detected in islets of both species, lactate accumulation is 6-fold higher in human islets. Human islets express LDH, with low-moderate LDHA expression and β cell-specific LDHB expression. LDHB inhibition amplifies LDHA-dependent lactate generation in mouse and human β cells and increases basal insulin release. Lastly, cis-instrument Mendelian randomization shows that low LDHB expression levels correlate with elevated fasting insulin in humans. Thus, LDHB limits lactate generation in β cells to maintain appropriate insulin release.

Leveraging cross-source heterogeneity to improve the performance of bulk gene expression deconvolution

A main limitation of bulk transcriptomic technologies is that individual measurements normally contain contributions from multiple cell populations, impeding the identification of cellular heterogeneity within diseased tissues. To extract cellular insights from existing large cohorts of bulk transcriptomic data, we present CSsingle, a novel method designed to accurately deconvolve bulk data into a predefined set of cell types using a scRNA-seq reference. Through comprehensive benchmark evaluations and analyses using diverse real data sets, we reveal the systematic bias inherent in existing methods, stemming from differences in cell size or library size. Our extensive experiments demonstrate that CSsingle exhibits superior accuracy and robustness compared to leading methods, particularly when dealing with bulk mixtures originating from cell types of markedly different cell sizes, as well as when handling bulk and single-cell reference data obtained from diverse sources. Our work provides an efficient and robust methodology for the integrated analysis of bulk and scRNA-seq data, facilitating various biological and clinical studies.

q-Diffusion leverages the full dimensionality of gene coexpression in single-cell transcriptomics

Unlocking the full dimensionality of single-cell RNA sequencing data (scRNAseq) is the next frontier to a richer, fuller understanding of cell biology. We introduce q-diffusion, a framework for capturing the coexpression structure of an entire library of genes, improving on state-of-the-art analysis tools. The method is demonstrated via three case studies. In the first, q-diffusion helps gain statistical significance for differential effects on patient outcomes when analyzing the CALGB/SWOG 80405 randomized phase III clinical trial, suggesting precision guidance for the treatment of metastatic colorectal cancer. Secondly, q-diffusion is benchmarked against existing scRNAseq classification methods using an in vitro PBMC dataset, in which the proposed method discriminates IFN-γ stimulation more accurately. The same case study demonstrates improvements in unsupervised cell clustering with the recent Tabula Sapiens human atlas. Finally, a local distributional segmentation approach for spatial scRNAseq, driven by q-diffusion, yields interpretable structures of human cortical tissue.

Distribution-Agnostic Deep Learning Enables Accurate Single-Cell Data Recovery and Transcriptional Regulation Interpretation

Single-cell RNA sequencing (scRNA-seq) is a robust method for studying gene expression at the single-cell level, but accurately quantifying genetic material is often hindered by limited mRNA capture, resulting in many missing expression values. Existing imputation methods rely on strict data assumptions, limiting their broader application, and lack reliable supervision, leading to biased signal recovery. To address these challenges, authors developed Bis, a distribution-agnostic deep learning model for accurately recovering missing sing-cell gene expression from multiple platforms. Bis is an optimal transport-based autoencoder model that can capture the intricate distribution of scRNA-seq data while addressing the characteristic sparsity by regularizing the cellular embedding space. Additionally, they propose a module using bulk RNA-seq data to guide reconstruction and ensure expression consistency. Experimental results show Bis outperforms other models across simulated and real datasets, showcasing superiority in various downstream analyses including batch effect removal, clustering, differential expression analysis, and trajectory inference. Moreover, Bis successfully restores gene expression levels in rare cell subsets in a tumor-matched peripheral blood dataset, revealing developmental characteristics of cytokine-induced natural killer cells within a head and neck squamous cell carcinoma microenvironment.

Single-cell sequencing: A promising approach for uncovering the characteristic of pancreatic islet cells in type 2 diabetes

Single-cell sequencing is a novel and rapidly advancing high-throughput technique that can be used to investigating genomics, transcriptomics, and epigenetics at a single-cell level. Currently, single-cell sequencing can not only be used to draw the pancreatic islet cells map and uncover the characteristics of cellular heterogeneity in type 2 diabetes, but can also be used to label and purify functional beta cells in pancreatic stem cells, improving stem cells and islet organoids therapies. In addition, this technology helps to analyze islet cell dedifferentiation and can be applied to the treatment of type 2 diabetes. In this review, we summarize the development and process of single-cell sequencing, describe the potential applications of single-cell sequencing in the field of type 2 diabetes, and discuss the prospects and limitations of single-cell sequencing to provide a new direction for exploring the pathogenesis of type 2 diabetes and finding therapeutic targets.

scAnnoX: an R package integrating multiple public tools for single-cell annotation

Background

Single-cell annotation plays a crucial role in the analysis of single-cell genomics data. Despite the existence of numerous single-cell annotation algorithms, a comprehensive tool for integrating and comparing these algorithms is also lacking.

Methods

This study meticulously investigated a plethora of widely adopted single-cell annotation algorithms. Ten single-cell annotation algorithms were selected based on the classification of either reference dataset-dependent or marker gene-dependent approaches. These algorithms included SingleR, Seurat, sciBet, scmap, CHETAH, scSorter, sc.type, cellID, scCATCH, and SCINA. Building upon these algorithms, we developed an R package named scAnnoX for the integration and comparative analysis of single-cell annotation algorithms.

Results

The development of the scAnnoX software package provides a cohesive framework for annotating cells in scRNA-seq data, enabling researchers to more efficiently perform comparative analyses among the cell type annotations contained in scRNA-seq datasets. The integrated environment of scAnnoX streamlines the testing, evaluation, and comparison processes among various algorithms. Among the ten annotation tools evaluated, SingleR, Seurat, sciBet, and scSorter emerged as top-performing algorithms in terms of prediction accuracy, with SingleR and sciBet demonstrating particularly superior performance, offering guidance for users. Interested parties can access the scAnnoX package at https://github.com/XQ-hub/scAnnoX.

A deep learning adversarial autoencoder with dynamic batching displays high performance in denoising and ordering scRNA-seq data

By providing high-resolution of cell-to-cell variation in gene expression, single-cell RNA sequencing (scRNA-seq) offers insights into cell heterogeneity, differentiating dynamics, and disease mechanisms. However, challenges such as low capture rates and dropout events can introduce noise in data analysis. Here, we propose a deep neural generative framework, the dynamic batching adversarial autoencoder (DB-AAE), which excels at denoising scRNA-seq datasets. DB-AAE directly captures optimal features from input data and enhances feature preservation, including cell type-specific gene expression patterns. Comprehensive evaluation on simulated and real datasets demonstrates that DB-AAE outperforms other methods in denoising accuracy and biological signal preservation. It also improves the accuracy of other algorithms in establishing pseudo-time inference. This study highlights DB-AAE's effectiveness and potential as a valuable tool for enhancing the quality and reliability of downstream analyses in scRNA-seq research.

Glucagon-producing α-cell transcriptional identity and reprogramming towards insulin production

β-Cell replacement by in situ reprogramming of non-β-cells is a promising diabetes therapy. Following the observation that near-total β-cell ablation in adult mice triggers the reprogramming of pancreatic α-, δ-, and γ-cells into insulin (INS)-producing cells, recent studies are delving deep into the mechanisms controlling adult α-cell identity. Systematic analyses of the α-cell transcriptome and epigenome have started to pinpoint features that could be crucial for maintaining α-cell identity. Using different transgenic and chemical approaches, significant advances have been made in reprogramming α-cells in vivo into INS-secreting cells in mice. The recent reprogramming of human α-cells in vitro is an important step forward that must now be complemented with a comprehensive molecular dissection of the mechanisms controlling α-cell identity.

RFX6 Maintains Gene Expression and Function of Adult Human Islet α-Cells

Mutations in the gene encoding the transcription factor regulatory factor X-box binding 6 (RFX6) are associated with human diabetes. Within pancreatic islets, RFX6 expression is most abundant in islet α-cells, and α-cell RFX6 expression is altered in diabetes. However, the roles of RFX6 in regulating gene expression, glucagon output, and other crucial human adult α-cell functions are not yet understood. We developed a method for selective genetic targeting of human α-cells and assessed RFX6-dependent α-cell function. RFX6 suppression with RNA interference led to impaired α-cell exocytosis and dysregulated glucagon secretion in vitro and in vivo. By contrast, these phenotypes were not observed with RFX6 suppression across all islet cells. Transcriptomics in α-cells revealed RFX6-dependent expression of genes governing nutrient sensing, hormone processing, and secretion, with some of these exclusively expressed in human α-cells. Mapping of RFX6 DNA-binding sites in primary human islet cells identified a subset of direct RFX6 target genes. Together, these data unveil RFX6-dependent genetic targets and mechanisms crucial for regulating adult human α-cell function.

ARTICLE HIGHLIGHTS

Decentralised, collaborative, and privacy-preserving machine learning for multi-hospital data

BACKGROUND

Machine Learning (ML) has demonstrated its great potential on medical data analysis. Large datasets collected from diverse sources and settings are essential for ML models in healthcare to achieve better accuracy and generalizability. Sharing data across different healthcare institutions or jurisdictions is challenging because of complex and varying privacy and regulatory requirements. Hence, it is hard but crucial to allow multiple parties to collaboratively train an ML model leveraging the private datasets available at each party without the need for direct sharing of those datasets or compromising the privacy of the datasets through collaboration.

METHODS

In this paper, we address this challenge by proposing Decentralized, Collaborative, and Privacy-preserving ML for Multi-Hospital Data (DeCaPH). This framework offers the following key benefits: (1) it allows different parties to collaboratively train an ML model without transferring their private datasets (i.e., no data centralization); (2) it safeguards patients' privacy by limiting the potential privacy leakage arising from any contents shared across the parties during the training process; and (3) it facilitates the ML model training without relying on a centralized party/server.

FINDINGS

We demonstrate the generalizability and power of DeCaPH on three distinct tasks using real-world distributed medical datasets: patient mortality prediction using electronic health records, cell-type classification using single-cell human genomes, and pathology identification using chest radiology images. The ML models trained with DeCaPH framework have less than 3.2% drop in model performance comparing to those trained by the non-privacy-preserving collaborative framework. Meanwhile, the average vulnerability to privacy attacks of the models trained with DeCaPH decreased by up to 16%. In addition, models trained with our DeCaPH framework achieve better performance than those models trained solely with the private datasets from individual parties without collaboration and those trained with the previous privacy-preserving collaborative training framework under the same privacy guarantee by up to 70% and 18.2% respectively.

INTERPRETATION

We demonstrate that the ML models trained with DeCaPH framework have an improved utility-privacy trade-off, showing DeCaPH enables the models to have good performance while preserving the privacy of the training data points. In addition, the ML models trained with DeCaPH framework in general outperform those trained solely with the private datasets from individual parties, showing that DeCaPH enhances the model generalizability.

FUNDING

This work was supported by the Natural Sciences and Engineering Research Council of Canada (NSERC, RGPIN-2020-06189 and DGECR-2020-00294), Canadian Institute for Advanced Research (CIFAR) AI Catalyst Grants, CIFAR AI Chair programs, Temerty Professor of AI Research and Education in Medicine, University of Toronto, Amazon, Apple, DARPA through the GARD project, Intel, Meta, the Ontario Early Researcher Award, and the Sloan Foundation. Resources used in preparing this research were provided, in part, by the Province of Ontario, the Government of Canada through CIFAR, and companies sponsoring the Vector Institute.

Metabolic control of adaptive β-cell proliferation by the protein deacetylase SIRT2

Selective and controlled expansion of endogenous β-cells has been pursued as a potential therapy for diabetes. Ideally, such therapies would preserve feedback control of β-cell proliferation to avoid excessive β-cell expansion and an increased risk of hypoglycemia. Here, we identified a regulator of β-cell proliferation whose inactivation results in controlled β-cell expansion: the protein deacetylase Sirtuin 2 (SIRT2). Sirt2 deletion in β-cells of mice increased β-cell proliferation during hyperglycemia with little effect in homeostatic conditions, indicating preservation of feedback control of β-cell mass. SIRT2 restrains proliferation of human islet β-cells cultured in glucose concentrations above the glycemic set point, demonstrating conserved SIRT2 function. Analysis of acetylated proteins in islets treated with a SIRT2 inhibitor revealed that SIRT2 deacetylates enzymes involved in oxidative phosphorylation, dampening the adaptive increase in oxygen consumption during hyperglycemia. At the transcriptomic level, Sirt2 inactivation has context-dependent effects on β-cells, with Sirt2 controlling how β-cells interpret hyperglycemia as a stress. Finally, we provide proof-of-principle that systemic administration of a GLP1-coupled Sirt2-targeting antisense oligonucleotide achieves β-cell selective Sirt2 inactivation and stimulates β-cell proliferation under hyperglycemic conditions. Overall, these studies identify a therapeutic strategy for increasing β-cell mass in diabetes without circumventing feedback control of β-cell proliferation.

Consideration of sex as a biological variable in diabetes research across twenty years

BACKGROUND

Sex differences exist in the risk of developing type 1 and type 2 diabetes, and in the risk of developing diabetes-associated complications. Sex differences in glucose homeostasis, islet and β cell biology, and peripheral insulin sensitivity have also been reported. Yet, we lack detailed information on the mechanisms underlying these differences, preventing the development of sex-informed therapeutic strategies for persons living with diabetes. To chart a path toward greater inclusion of biological sex as a variable in diabetes research, we first need a detailed assessment of common practices in the field.

METHODS

We developed a scoring system to evaluate the inclusion of biological sex in manuscripts published in Diabetes, a journal published by the American Diabetes Association. We chose Diabetes as this journal focuses solely on diabetes and diabetes-related research, and includes manuscripts that use both clinical and biomedical approaches. We scored papers published across 3 years within a 20-year period (1999, 2009, 2019), a timeframe that spans the introduction of funding agency and journal policies designed to improve the consideration of biological sex as a variable.

RESULTS

Our analysis showed fewer than 15% of papers used sex-based analysis in even one figure across all study years, a trend that was reproduced across journal-defined categories of diabetes research (e.g., islet studies, signal transduction). Single-sex studies accounted for approximately 40% of all manuscripts, of which > 87% used male subjects only. While we observed a modest increase in the overall inclusion of sex as a biological variable during our study period, our data highlight significant opportunities for improvement in diabetes research practices. We also present data supporting a positive role for journal policies in promoting better consideration of biological sex in diabetes research.

CONCLUSIONS

Our analysis provides significant insight into common practices in diabetes research related to the consideration of biological sex as a variable. Based on our analysis we recommend ways that diabetes researchers can improve inclusion of biological sex as a variable. In the long term, improved practices will reveal sex-specific mechanisms underlying diabetes risk and complications, generating knowledge to enable the development of sex-informed prevention and treatment strategies.

The Role of G Protein-Coupled Receptors and Receptor Kinases in Pancreatic β-Cell Function and Diabetes

Type 2 diabetes (T2D) mellitus has emerged as a major global health concern that has accelerated in recent years due to poor diet and lifestyle. Afflicted individuals have high blood glucose levels that stem from the inability of the pancreas to make enough insulin to meet demand. Although medication can help to maintain normal blood glucose levels in individuals with chronic disease, many of these medicines are outdated, have severe side effects, and often become less efficacious over time, necessitating the need for insulin therapy. G protein-coupled receptors (GPCRs) regulate many physiologic processes, including blood glucose levels. In pancreatic β cells, GPCRs regulate β-cell growth, apoptosis, and insulin secretion, which are all critical in maintaining sufficient β-cell mass and insulin output to ensure euglycemia. In recent years, new insights into the signaling of incretin receptors and other GPCRs have underscored the potential of these receptors as desirable targets in the treatment of diabetes. The signaling of these receptors is modulated by GPCR kinases (GRKs) that phosphorylate agonist-activated GPCRs, marking the receptor for arrestin binding and internalization. Interestingly, genome-wide association studies using diabetic patient cohorts link the GRKs and arrestins with T2D. Moreover, recent reports show that GRKs and arrestins expressed in the β cell serve a critical role in the regulation of β-cell function, including β-cell growth and insulin secretion in both GPCR-dependent and -independent pathways. In this review, we describe recent insights into GPCR signaling and the importance of GRK function in modulating β-cell physiology. SIGNIFICANCE STATEMENT: Pancreatic β cells contain a diverse array of G protein-coupled receptors (GPCRs) that have been shown to improve β-cell function and survival, yet only a handful have been successfully targeted in the treatment of diabetes. This review discusses recent advances in our understanding of β-cell GPCR pharmacology and regulation by GPCR kinases while also highlighting the necessity of investigating islet-enriched GPCRs that have largely been unexplored to unveil novel treatment strategies.

A novel batch-effect correction method for scRNA-seq data based on Adversarial Information Factorization

Single-cell RNA sequencing (scRNA-seq) technology produces an unprecedented resolution at the level of a unique cell, raising great hopes in medicine. Nevertheless, scRNA-seq data suffer from high variations due to the experimental conditions, called batch effects, preventing any aggregated downstream analysis. Adversarial Information Factorization provides a robust batch-effect correction method that does not rely on prior knowledge of the cell types nor a specific normalization strategy while being adapted to any downstream analysis task. It compares to and even outperforms state-of-the-art methods in several scenarios: low signal-to-noise ratio, batch-specific cell types with few cells, and a multi-batches dataset with imbalanced batches and batch-specific cell types. Moreover, it best preserves the relative gene expression between cell types, yielding superior differential expression analysis results. Finally, in a more complex setting of a Leukemia cohort, our method preserved most of the underlying biological information for each patient while aligning the batches, improving the clustering metrics in the aggregated dataset.

STING signaling in islet macrophages impairs insulin secretion in obesity

The innate immune regulator stimulator of interferon genes (STING) mediates self-DNA sensing and leads to the induction of type I interferons and inflammatory cytokines, which promotes the progression of various inflammatory and autoimmune diseases. Innate immune system plays a critical role in regulating obesity-induced islet dysfunction, whereas the potential effect of STING signaling is not fully understood. Here, we demonstrate that STING is mainly expressed and activated in islet macrophages upon high-fat diet (HFD) feeding. Sting-/- alleviates HFD-induced islet inflammation by inhibiting the expression of pro-inflammatory cytokines and the infiltration of macrophages. Mechanically, palmitic acid incubation promotes mitochondrial DNA leakage into the cytosol and subsequently activates STING pathway in macrophages. Additionally, STING activation in macrophages impairs glucose-stimulated insulin secretion by mediating the engulfment of β cell insulin secretory granules. Pharmacologically inhibiting STING activation enhances insulin secretion to control hyperglycemia. Together, our results reveal a regulatory mechanism in controlling the islet inflammation and insulin secretion in diet--induced obesity and suggest that selective blocking of the STING activation may be a promising strategy for treating type 2 diabetes.

Predicting cell-type specific disease genes of diabetes with the biological network

Type 2 diabetes (T2D) is a chronic condition that can lead to significant harm, such as heart disease, kidney disease, nerve damage, and blindness. Although T2D-related genes have been identified through Genome-wide association studies (GWAS) and various computational methods, the biological mechanism of T2D at the cell type level remains unclear. Exploring cell type-specific genes related to T2D is essential to understand the cellular mechanisms underlying the disease. To address this issue, we introduce DiGCellNet (predicting Disease Genes with Cell type specificity based on biological Networks), a model that integrates graph convolutional network (GCN) and multi-task learning (MTL) to predict T2D-associated cell type-specific genes based on the biological network. Our work represents the first attempt to predict cell type-specific disease genes using GCN and MTL. We evaluate our approach by predicting genes specific to four cell types and demonstrate that the proposed DiGCellNet outperforms other models that combine node embeddings with traditional machine learning algorithms. Moreover, DiGCellNet successfully identifies CALM1 as a gene specific to beta cell type in T2D cases, and this association is confirmed using an independent dataset. The code is available at https://github.com/23AIBox/23AIBox-DiGCellNet.

A New Graph Autoencoder-Based Consensus-Guided Model for scRNA-seq Cell Type Detection

Single-cell RNA sequencing (scRNA-seq) technology is famous for providing a microscopic view to help capture cellular heterogeneity. This characteristic has advanced the field of genomics by enabling the delicate differentiation of cell types. However, the properties of single-cell datasets, such as high dropout events, noise, and high dimensionality, are still a research challenge in the single-cell field. To utilize single-cell data more efficiently and to better explore the heterogeneity among cells, a new graph autoencoder (GAE)-based consensus-guided model (scGAC) is proposed in this article. The data are preprocessed into multiple top-level feature datasets. Then, feature learning is performed by using GAEs to generate new feature matrices, followed by similarity learning based on distance fusion methods. The learned similarity matrices are fed back to the GAEs to guide their feature learning process. Finally, the abovementioned steps are iterated continuously to integrate the final consistent similarity matrix and perform other related downstream analyses. The scGAC model can accurately identify critical features and effectively preserve the internal structure of the data. This can further improve the accuracy of cell type identification.

Identification of structural motifs critical for human G6PC2 function informed by sequence analysis and an AlphaFold2-predicted model

G6PC2 encodes a glucose-6-phosphatase (G6Pase) catalytic subunit, primarily expressed in pancreatic islet β cells, which modulates the sensitivity of insulin secretion to glucose and thereby regulates fasting blood glucose (FBG). Mutational analyses were conducted to validate an AlphaFold2 (AF2)-predicted structure of human G6PC2 in conjunction with a novel method to solubilize and purify human G6PC2 from a heterologous expression system. These analyses show that residues forming a predicted intramolecular disulfide bond are essential for G6PC2 expression and that residues forming part of a type 2 phosphatidic acid phosphatase (PAP2) motif are critical for enzyme activity. Additional mutagenesis shows that residues forming a predicted substrate cavity modulate enzyme activity and substrate specificity and residues forming a putative cholesterol recognition amino acid consensus (CRAC) motif influence protein expression or enzyme activity. This CRAC motif begins at residue 219, the site of a common G6PC2 non-synonymous single-nucleotide polymorphism (SNP), rs492594 (Val219Leu), though the functional impact of this SNP is disputed. In microsomal membrane preparations, the L219 variant has greater activity than the V219 variant, but this difference disappears when G6PC2 is purified in detergent micelles. We hypothesize that this was due to a differential association of the two variants with cholesterol. This concept was supported by the observation that the addition of cholesteryl hemi-succinate to the purified enzymes decreased the Vmax of the V219 and L219 variants ∼8-fold and ∼3 fold, respectively. We anticipate that these observations should support the rational development of G6PC2 inhibitors designed to lower FBG.

Anti-correlated feature selection prevents false discovery of subpopulations in scRNAseq

While sub-clustering cell-populations has become popular in single cell-omics, negative controls for this process are lacking. Popular feature-selection/clustering algorithms fail the null-dataset problem, allowing erroneous subdivisions of homogenous clusters until nearly each cell is called its own cluster. Using real and synthetic datasets, we find that anti-correlated gene selection reduces or eliminates erroneous subdivisions, increases marker-gene selection efficacy, and efficiently scales to millions of cells.

Comparative and integrative single cell analysis reveals new insights into the transcriptional immaturity of stem cell-derived β cells

Diabetes cell replacement therapy has the potential to be transformed by human pluripotent stem cell-derived β cells (SC-β cells). However, the precise identity of SC-β cells in relationship to primary fetal and adult β-cells remains unclear. Here, we used single-cell sequencing datasets to characterize the transcriptional identity of islets from in vitro differentiation, fetal islets, and adult islets. Our analysis revealed that SC-β cells share a core β-cell transcriptional identity with human adult and fetal β-cells, however SC-β cells possess a unique transcriptional profile characterized by the persistent expression and activation of progenitor and neural-biased gene networks. These networks are present in SC-β cells, irrespective of the derivation protocol used. Notably, fetal β-cells also exhibit this neural signature at the transcriptional level. Our findings offer insights into the transcriptional identity of SC-β cells and underscore the need for further investigation of the role of neural transcriptional networks in their development.

Identification of type 2 diabetes- and obesity-associated human β-cells using deep transfer learning

Diabetes affects >10% of adults worldwide and is caused by impaired production or response to insulin, resulting in chronic hyperglycemia. Pancreatic islet β-cells are the sole source of endogenous insulin and our understanding of β-cell dysfunction and death in type 2 diabetes (T2D) is incomplete. Single-cell RNA-seq data supports heterogeneity as an important factor in β-cell function and survival. However, it is difficult to identify which β-cell phenotypes are critical for T2D etiology and progression. Our goal was to prioritize specific disease-related β-cell subpopulations to better understand T2D pathogenesis and identify relevant genes for targeted therapeutics. To address this, we applied a deep transfer learning tool, DEGAS, which maps disease associations onto single-cell RNA-seq data from bulk expression data. Independent runs of DEGAS using T2D or obesity status identified distinct β-cell subpopulations. A singular cluster of T2D-associated β-cells was identified; however, β-cells with high obese-DEGAS scores contained two subpopulations derived largely from either non-diabetic or T2D donors. The obesity-associated non-diabetic cells were enriched for translation and unfolded protein response genes compared to T2D cells. We selected DLK1 for validation by immunostaining in human pancreas sections from healthy and T2D donors. DLK1 was heterogeneously expressed among β-cells and appeared depleted from T2D islets. In conclusion, DEGAS has the potential to advance our holistic understanding of the β-cell transcriptomic phenotypes, including features that distinguish β-cells in obese non-diabetic or lean T2D states. Future work will expand this approach to additional human islet omics datasets to reveal the complex multicellular interactions driving T2D.

scAMAC: self-supervised clustering of scRNA-seq data based on adaptive multi-scale autoencoder

Cluster assignment is vital to analyzing single-cell RNA sequencing (scRNA-seq) data to understand high-level biological processes. Deep learning-based clustering methods have recently been widely used in scRNA-seq data analysis. However, existing deep models often overlook the interconnections and interactions among network layers, leading to the loss of structural information within the network layers. Herein, we develop a new self-supervised clustering method based on an adaptive multi-scale autoencoder, called scAMAC. The self-supervised clustering network utilizes the Multi-Scale Attention mechanism to fuse the feature information from the encoder, hidden and decoder layers of the multi-scale autoencoder, which enables the exploration of cellular correlations within the same scale and captures deep features across different scales. The self-supervised clustering network calculates the membership matrix using the fused latent features and optimizes the clustering network based on the membership matrix. scAMAC employs an adaptive feedback mechanism to supervise the parameter updates of the multi-scale autoencoder, obtaining a more effective representation of cell features. scAMAC not only enables cell clustering but also performs data reconstruction through the decoding layer. Through extensive experiments, we demonstrate that scAMAC is superior to several advanced clustering and imputation methods in both data clustering and reconstruction. In addition, scAMAC is beneficial for downstream analysis, such as cell trajectory inference. Our scAMAC model codes are freely available at https://github.com/yancy2024/scAMAC.

scDOT: enhancing single-cell RNA-Seq data annotation and uncovering novel cell types through multi-reference integration

The proliferation of single-cell RNA-seq data has greatly enhanced our ability to comprehend the intricate nature of diverse tissues. However, accurately annotating cell types in such data, especially when handling multiple reference datasets and identifying novel cell types, remains a significant challenge. To address these issues, we introduce Single Cell annotation based on Distance metric learning and Optimal Transport (scDOT), an innovative cell-type annotation method adept at integrating multiple reference datasets and uncovering previously unseen cell types. scDOT introduces two key innovations. First, by incorporating distance metric learning and optimal transport, it presents a novel optimization framework. This framework effectively learns the predictive power of each reference dataset for new query data and simultaneously establishes a probabilistic mapping between cells in the query data and reference-defined cell types. Secondly, scDOT develops an interpretable scoring system based on the acquired probabilistic mapping, enabling the precise identification of previously unseen cell types within the data. To rigorously assess scDOT's capabilities, we systematically evaluate its performance using two diverse collections of benchmark datasets encompassing various tissues, sequencing technologies and diverse cell types. Our experimental results consistently affirm the superior performance of scDOT in cell-type annotation and the identification of previously unseen cell types. These advancements provide researchers with a potent tool for precise cell-type annotation, ultimately enriching our understanding of complex biological tissues.

scCorrector: a robust method for integrating multi-study single-cell data

The advent of single-cell sequencing technologies has revolutionized cell biology studies. However, integrative analyses of diverse single-cell data face serious challenges, including technological noise, sample heterogeneity, and different modalities and species. To address these problems, we propose scCorrector, a variational autoencoder-based model that can integrate single-cell data from different studies and map them into a common space. Specifically, we designed a Study Specific Adaptive Normalization for each study in decoder to implement these features. scCorrector substantially achieves competitive and robust performance compared with state-of-the-art methods and brings novel insights under various circumstances (e.g. various batches, multi-omics, cross-species, and development stages). In addition, the integration of single-cell data and spatial data makes it possible to transfer information between different studies, which greatly expand the narrow range of genes covered by MERFISH technology. In summary, scCorrector can efficiently integrate multi-study single-cell datasets, thereby providing broad opportunities to tackle challenges emerging from noisy resources.