Single-cell Multi-omics Sequencing Research Articles

ScDRMAE: Integrating Masked Autoencoder with Residual Attention Networks to Leverage Omics Feature Dependencies for Accurate Cell Clustering.

Cell clustering is foundational for analyzing the heterogeneity of biological tissues using single-cell sequencing data. With the maturation of single-cell multi-omics sequencing technologies, we can integrate multiple omics data to perform cell clustering, thereby overcoming the limitations of insufficient information from single omics data. Existing methods for cell clustering often only consider the differences in data patterns during the analysis of multi-omics data, but the dependencies between omics features of different cell types also significantly influence cell clustering. Moreover, the high dropout rates in scRNA-seq and scATAC-seq data can impact the performance of cell clustering. We propose a cell clustering model based on a masked autoencoder, scDRMAE. Utilizing a masking mechanism, scDRMAE effectively learns the relationships between different features and imputes false zeros caused by dropout events. To differentiate the importance of various omics data in cell clustering, we dynamically adjust the weights of different omics data through an attention mechanism. Finally, we use the K-means algorithm for cluster analysis of the fused multi-omics data. On commonly used sets of 15 multi-omics datasets, our method demonstrates superior cell clustering performance on multiple metrics compared to other computational methods. Additionally, when datasets exhibit varying degrees of dropout noise, our method shows better performance and stronger stability on multiple metrics compared to other methods. Moreover, by analyzing the cell clusters classified by scDRMAE, we identified several biologically significant biomarkers that have been validated, further confirming the effectiveness of scDRMAE in cell clustering from a biological perspective. Supplementary data are available at Bioinformatics online.

Simultaneous profiling of RNA isoforms and chromatin accessibility of single cells of human retinal organoids

Single-cell multi-omics sequencing is a powerful approach to analyze complex mechanisms underlying neuronal development and regeneration. However, current methods lack the ability to simultaneously profile RNA alternative splicing and chromatin accessibility at the single-cell level. We develop a technique, single-cell RNA isoform and chromatin accessibility sequencing (scRICA-seq), which demonstrates higher sensitivity and cost-effectiveness compared to existing methods. scRICA-seq can profile both isoforms and chromatin accessibility for up to 10,000 single cells in a single run. Applying this method to human retinal organoids, we construct a multi-omic cell atlas and reveal associations between chromatin accessibility, isoform expression of fate-determining factors, and alternative splicing events in their binding sites. This study provides insights into integrating epigenetics, transcription, and RNA splicing to elucidate the mechanisms underlying retinal neuronal development and fate determination.

O-215 Copper oxide nanoparticles (CuONPs) impairs mitophagy and promotes DNA methylation to suppress mouse embryonic pluripotency

Abstract Study question Does exposure to copper oxide nanoparticle (CuONPs) have adverse effects on early embryonic development? What is the potential mechanism? Summary answer CuONPs exposure can damage mitophagy and lead to α-KG reduction, resulting DNA hypermethylation, thereby reducing the pluripotency of pre-implantation embryos in mice. What is known already Environmental nanomaterials contamination is a great concern for organisms including human. Exposure to nanomaterials have been associated with fecundity and fertility in couples conceiving via assisted reproduction technology. Among these bio-effects, nanomaterials induced epigenetic changes are of particular interest. Copper oxide nanoparticles (CuONPs) are widely used in a huge range of applications which might pose potential risk to organisms. Recent studies reported that CuONPs exposure resulted in transgenerational toxicity in Caenorhabditis elegans associated with possible epigenetic regulation. However, whether CuONPs exposure affect mammalian embryonic development, and if so the underlying mechanisms, are unclear. Study design, size, duration For the human experiments, we measured the copper ion content in women’s urine, and assessed the expression of pluripotent genes in 3PN embryos treated with CuONPs. For the embryo experiments, we collected zygotes from ICR mice and cultured them in KSOM medium with CuONPs (10 μg/mL) until blastocyst stage. For the cell experiments, we treated mouse embryonic stem cells (mESCs) with CuONPs (2 μg/mL) for four days to detect cell proliferation and pluripotency. Participants/materials, setting, methods The ICP-MS method was used to detect copper ion content in women’s urine. Time–lapse incubator was used to document the developmental potential of embryos. To quantify the metabolites of embryo, we employed metabolomics. Single-cell multiomics sequencing that enables profiling of genome-wide chromatin accessibility, DNA methylation and RNA expression were performed to detect genes expression and epigenetic modification. We used real-time PCR, immunofluorescence, and western blot to measure the expression of pluripotency related genes. Main results and the role of chance In human, we discovered a significant negative relationship between the urine’s copper ion content and clinical pregnancy rate. In embryo experiments, CuONPs exposure has been shown to have adverse effects on embryo developmental potential, resulting morula and blastocyst transformation arrest, as well as reduction in blastocyst rate and live birth rate. Furthermore, CuONPs treatment led to a reduction in blastocyst quality, indicating a decrease in blastocyst cell count, particularly in ICM cells. CuONPs also inhibited the pluripotency of blastocysts and mESCs, showing down-regulated expression of pluripotent genes. Mechanistically, CuONPs aggregated in lysosomes and decreased the mitophagic flux of embryos, which caused damaged mitochondria to accumulate and ultimately caused mitochondrial dysfunction. Notably, in embryos treated with CuONPs, there was a decrease in α-ketoglutarate (α-KG) due to the disruption of the tricarboxylic acid (TCA) cycle. α-KG is an essential cofactor for α-KG dependent dioxygenases such as the ten-eleven translocation (TET) family of DNA demethylase. The lack of α-KG in CuONPs-treated embryos led to DNA hypermethylation, which reduced chromatin accessibility, and suppressed the expression of pluripotent genes like sox2, klf2, and klf4. The poor pluripotency ultimately leads to abnormal embryonic development in CuONPs-treated embryos. Limitations, reasons for caution The molecular mechanism by which CuONPs arrest the development of pre-implantation embryos remains to be determined. In addition, future work should be extended to human embryos to determine whether this mechanism is conserved between mice and humans. Wider implications of the findings The findings of this work indicated the relationship between CuONPs and epigenetic modification. The results will support a more thorough understanding into the biosafety of CuONPs, and offer fresh perspectives on the potential uses of nanomaterials in the field of reproductive medicine in the future. Trial registration number Not Applicable

Single-cell analysis of anti-BCMA CAR T cell therapy in patients with central nervous system autoimmunity.

Chimeric antigen receptor (CAR) T cell immunotherapy for the treatment of neurological autoimmune diseases is promising, but CAR T cell kinetics and immune alterations after treatment are poorly understood. Here, we performed single-cell multi-omics sequencing of paired cerebrospinal fluid (CSF) and blood samples from patients with neuromyelitis optica spectrum disorder (NMOSD) treated with anti-B cell maturation antigen (BCMA) CAR T cells. Proliferating cytotoxic-like CD8+ CAR T cell clones were identified as the main effectors in autoimmunity. Anti-BCMA CAR T cells with enhanced features of chemotaxis efficiently crossed the blood-CSF barrier, eliminated plasmablasts and plasma cells in the CSF, and suppressed neuroinflammation. The CD44-expressing early memory phenotype in infusion products was potentially associated with CAR T cell persistence in autoimmunity. Moreover, CAR T cells from patients with NMOSD displayed distinctive features of suppressed cytotoxicity compared with those from hematological malignancies. Thus, we provide mechanistic insights into CAR T cell function in patients with neurological autoimmune disease.

Clustering single-cell multi-omics data via graph regularized multi-view ensemble learning.

Single-cell clustering plays a crucial role in distinguishing between cell types, facilitating the analysis of cell heterogeneity mechanisms. While many existing clustering methods rely solely on gene expression data obtained from single-cell RNA sequencing techniques to identify cell clusters, the information contained in mono-omic data is often limited, leading to suboptimal clustering performance. The emergence of single-cell multi-omics sequencing technologies enables the integration of multiple omics data for identifying cell clusters, but how to integrate different omics data effectively remains challenging. In addition, designing a clustering method that performs well across various types of multi-omics data poses a persistent challenge due to the data's inherent characteristics. In this paper, we propose a graph-regularized multi-view ensemble clustering (GRMEC-SC) model for single-cell clustering. Our proposed approach can adaptively integrate multiple omics data and leverage insights from multiple base clustering results. We extensively evaluate our method on five multi-omics datasets through a series of rigorous experiments. The results of these experiments demonstrate that our GRMEC-SC model achieves competitive performance across diverse multi-omics datasets with varying characteristics. Implementation of GRMEC-SC, along with examples, can be found on the GitHub repository: https://github.com/polarisChen/GRMEC-SC.

Abstract 5540: Single cell multiomic analysis of human brain metastasis reveals conserved dormancy population

Abstract This study explores the complex realm of brain metastasis, pivotal in crafting effective anti-tumor strategies. Leveraging single-cell multi-omics sequencing, we profiled the transcriptomes and chromatin accessibility landscapes of tumor cells and their surrounding microenvironment in human brain metastases that originated from across a spectrum of primary cancers. Our investigations have identified distinct, yet uniformly present, tumor cell subpopulations, each defined by unique gene expression patterns, chromatin accessibility, and the interactions between tumor and stromal cells. Notably, one such subpopulation exhibits a dormancy signature, characterized by enhanced interferon response, oxidative phosphorylation, and upregulated HLA class molecules, alongside alterations in circadian rhythms. Our trajectory analysis charts a course from these dormant cells to rapidly proliferating ones, with notable shifts in circadian rhythm genes, underscoring their potential role in brain metastasis. This aligns with recent findings that disruptions in circadian rhythms can modify stem cell characteristics in glioblastomas. This insight led us to focus on a specific circadian rhythm gene, highly active in the dormant cells and located on the HER2 amplicon—an area linked to high brain metastasis relapse rates in patients. We've developed inducible models for both overexpression and knockdown to scrutinize this gene's impact in her2 positive breast cancer cell lines, examining gene expression and proliferation in vitro. Complementing this, we are conducting spatial transcriptomic analysis to evaluate the microenvironmental regulation of tumor dormancy in human brain metastases. This multi-omics approach reveals a nuanced regulatory landscape of brain metastasis, highlighting a novel and often overlooked factor in metastasis progression and brain tumor relapse: the interplay between circadian rhythms and tumor dormancy. This discovery not only deepens our understanding of brain metastases but also opens potential new avenues for therapeutic interventions. Citation Format: Remi Klotz, Yiru Wang, Anu Sunkara, Frank Attenelo, Min Yu. Single cell multiomic analysis of human brain metastasis reveals conserved dormancy population [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 5540.

Abstract 5862: Retrotransposable elements mediate the drug-tolerant persistence in claudin-low breast cancer chemo-treatment

Abstract The emergence of acquired drug resistance through treatment remains a critical threat to efficient cancer therapies, which most often lead to relapse and metastasis. The development of resistance is an evolutionary adaptation process that heavily depends on tumor heterogeneity (underpinned by cellular plasticity). Therefore, a longitudinal deconvolution of treatment adaptation is critical to uncover the driving-force mechanisms underpinning resistance adaptation. In the clinic, chemotherapy remains the mainstream treatment for TNBC, and one of the primary chemo-agents is doxorubicin. Although the initial responsive rate of doxorubicin-based chemotherapies is up to 70%, it is well recognized that TNBC cells usually generate an adaptive response and result in acquired drug-resistance and multi-drug resistant phenotypes. Although numerous mechanisms of chemo-resistance evolution have been proposed, most of these studies focused on the initial and terminal states, from which we could not deconvolute the adaptation routes and might miss the transient fate-switching events. We hypothesize that the claudin-low TNBC chemo-resistant cells may derive from the early-stage reversible chemo-tolerant persistent state, and early-stage state-switching mediated by epigenetic landscape reprogramming might determine the adaptation through treatment. We developed an in vitro “pulsing-treatment” model to test the hypothesis, which could mimic the clinical treatment and provide therapeutically relevant insights into the acute drug-induced stress response and regrowth. Unexpectedly, we found that the human endogenous retrovirus-like elements (HERVs) show an aberrant bursting in the adaptation. To further interrogate the adaptation, we did the longitudinal single-cell multi-omics sequencing. With a novel bioinformatic workflow, we integrated the HERV’s expression information with protein-coding genes profiling and chromatin accessibility. The preliminary result indicated that a subpopulation of HERV-high cells might transiently mediate the persistence of cancer cells at the early stage of the treatment, which coupled with whole transcriptomic and chromatin landscape reprogramming, and molecular subtype switching. In this study, we employed a new strategy to investigate the longitudinal adaptation route through treatment and uncover a non-canonical element, which shed new light on drug resistance research and novel target screening. Citation Format: Zijian Zhang, Yiyang Wang, Xinlong Luo, Xiaomei Zhan, Jun Ding, Tao P. Wu. Retrotransposable elements mediate the drug-tolerant persistence in claudin-low breast cancer chemo-treatment [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 5862.

MoSCminer: a cell subtype classification framework based on the attention neural network integrating the single-cell multi-omics dataset on the cloud.

Single-cell omics sequencing has rapidly advanced, enabling the quantification of diverse omics profiles at a single-cell resolution. To facilitate comprehensive biological insights, such as cellular differentiation trajectories, precise annotation of cell subtypes is essential. Conventional methods involve clustering cells and manually assigning subtypes based on canonical markers, a labor-intensive and expert-dependent process. Hence, an automated computational prediction framework is crucial. While several classification frameworks for predicting cell subtypes from single-cell RNA sequencing datasets exist, these methods solely rely on single-omics data, offering insights at a single molecular level. They often miss inter-omic correlations and a holistic understanding of cellular processes. To address this, the integration of multi-omics datasets from individual cells is essential for accurate subtype annotation. This article introduces moSCminer, a novel framework for classifying cell subtypes that harnesses the power of single-cell multi-omics sequencing datasets through an attention-based neural network operating at the omics level. By integrating three distinct omics datasets-gene expression, DNA methylation, and DNA accessibility-while accounting for their biological relationships, moSCminer excels at learning the relative significance of each omics feature. It then transforms this knowledge into a novel representation for cell subtype classification. Comparative evaluations against standard machine learning-based classifiers demonstrate moSCminer's superior performance, consistently achieving the highest average performance on real datasets. The efficacy of multi-omics integration is further corroborated through an in-depth analysis of the omics-level attention module, which identifies potential markers for cell subtype annotation. To enhance accessibility and scalability, moSCminer is accessible as a user-friendly web-based platform seamlessly connected to a cloud system, publicly accessible at http://203.252.206.118:5568. Notably, this study marks the pioneering integration of three single-cell multi-omics datasets for cell subtype identification.

Single-cell multi-omics profiling of human preimplantation embryos identifies cytoskeletal defects during embryonic arrest.

Human in vitro fertilized embryos exhibit low developmental capabilities, and the mechanisms that underlie embryonic arrest remain unclear. Here using a single-cell multi-omics sequencing approach, we simultaneously analysed alterations in the transcriptome, chromatin accessibility and the DNA methylome in human embryonic arrest due to unexplained reasons. Arrested embryos displayed transcriptome disorders, including a distorted microtubule cytoskeleton, increased genomic instability and impaired glycolysis, which were coordinated with multiple epigenetic reprogramming defects. We identified Aurora A kinase (AURKA) repression as a cause of embryonic arrest. Mechanistically, arrested embryos induced through AURKA inhibition resembled the reprogramming abnormalities of natural embryonic arrest in terms of the transcriptome, the DNA methylome, chromatin accessibility and H3K4me3 modifications. Mitosis-independent sequential activation of the zygotic genome in arrested embryos showed that YY1 contributed to human major zygotic genome activation. Collectively, our study decodes the reprogramming abnormalities and mechanisms of human embryonic arrest and the key regulators of zygotic genome activation.

SMARTdb: An Integrated Database for Exploring Single-cell Multi-omics Data of Reproductive Medicine.

Single-cell multi-omics sequencing has greatly accelerated reproductive research in recent years, and the data are continually growing. However, utilizing these data resources is challenging for wet-lab researchers. A comprehensive platform for exploring single-cell multi-omics data related to reproduction is urgently needed. Here, we introduce the single-cell multi-omics atlas of reproduction (SMARTdb), an integrative and user-friendly platform for exploring molecular dynamics of reproductive development, aging, and disease, which covers multi-omics, multi-species, and multi-stage data. We curated and analyzed single-cell transcriptomic and epigenomic data of over 2.0 million cells from 6 species across the entire lifespan. A series of powerful functionalities are provided, such as "Query gene expression", "DIY expression plot", "DNA methylation plot", and "Epigenome browser". With SMARTdb, we found that the male germ cell-specific expression pattern of RPL39L and RPL10L is conserved between human and other model animals. Moreover, DNA hypomethylation and open chromatin may collectively regulate the specific expression pattern of RPL39L in both male and female germ cells. In summary, SMARTdb is a powerful platform for convenient data mining and gaining novel insights into reproductive development, aging, and disease. SMARTdb is publicly available at https://smart-db.cn.

Linking DNA Damage Response to Ferroptosis for Leukemia Stem Cell Eradication

Relapse in acute myeloid leukemia (AML) is associated with an unfavorable prognosis, primarily due to the poor response to conventional chemotherapy, which has shown disappointing overall outcomes over the past 30 years. The main culprits behind leukemogenesis and therapy resistance are leukemia stem cells (LSCs), characterized by their unique features of self-renewal and evasion from cell death. In our recent work, we have identified a critical self-renewal pathway, RSPO3-LGR4/p-PKAc, which exhibits aberrant activation in relapsed or refractory AML (Salik et al. Cancer Cell, 2020). Blocking this pathway holds the potential to eradicate resistant LSCs and treat relapsed AML effectively. In this study, our data directly link the self-renewal pathway to ferroptosis resistance through activation of antioxidant defense, which confers protection on AML LSCs against detrimental effects of oxidative stress. We also found that aberrant activation of RSPO3-LGR4/p-PKAc suppressed the DNA damage response (DDR) pathway consequently leading to genomic instability, DNA repair defects, increased stemness and AML aggressiveness. Our whole-genome sequencing and knockout mouse studies showed that deletion of the key DDR gene caused substantial mutations on multiple pathways including DNA damage repair (e.g., Slf1) and self-renewal (e.g., Hox cluster genes), contributing to a more aggressive and therapy-resistant phenotype in AML initiation and progression. Single cell multi-omics sequencing and CRISPR/Cas9 knockout mouse models uncovered essential antioxidant enzymes as downstream targets of DDR pathway inhibition in AML patient-derived LSCs. These antioxidant enzymes induced by DDR pathway inhibition had protective effects against oxidative attack due to the ability to decompose reactive oxygen species (ROS) and detoxify excess iron. ShRNA-mediated stable knockdown of antioxidant enzymes impaired LSCs by increasing accumulation of iron and ROS and sensitized LSCs to ferroptosis (an iron-dependent and ROS-reliant cell death). Our analysis of AML dataset from TCGA (The Cancer Genome Atlas, N Engl J Med 2013) displayed a positive correlation of high expression of antioxidant genes with poor prognosis in a cohort of 172 AML patients (p<0.019), indicating the clinical importance of antioxidant dysregulation in cancer. Collectively, we report here a functional relationship of DDR pathway inhibition and antioxidant defense mechanism that is regulated by RSPO3-LGR4/p-PKAc self-renewal pathway and contributes to ferroptosis resistance in AML LSCs. Targeting essential pathway components enables LSCs to become sensitive to oxidative damage and ferroptosis and perturbs leukemia progression, presenting a promising new therapeutic approach for relapsed or refractory AML.

Single-cell multi-omics sequencing reveals chromosome copy number inconsistency between trophectoderm and inner cell mass in human reconstituted embryos after spindle transfer.

Is the chromosome copy number of the trophectoderm (TE) of a human reconstituted embryos after spindle transfer (ST) representative of the inner cell mass (ICM)? Single-cell multi-omics sequencing revealed that ST blastocysts have a higher proportion of cell lineages exhibiting intermediate mosaicism than conventional ICSI blastocysts, and that the TE of ST blastocysts does not represent the chromosome copy number of ICM. Preimplantation genetic testing for aneuploidy (PGT-A) assumes that TE biopsies are representative of the ICM, but the TE and ICM originate from different cell lineages, and concordance between TE and ICM is not well-studied, especially in ST embryos. We recruited 30 infertile women who received treatment at our clinic and obtained 45 usable blastocysts (22 from conventional ICSI and 23 reconstituted embryos after ST). We performed single-cell multi-omics sequencing on all blastocysts to predict and verify copy number variations (CNVs) in each cell. We determined the chromosome copy number of each embryo by analysing the proportion of abnormal cells in each blastocyst. We used the Bland-Altman concordance and the Kappa test to evaluate the concordance between TE and ICM in the both groups. The study was conducted at a public tertiary hospital in China, where all the embryo operations, including oocytes retrieval, ST, and ICSI, were performed in the embryo laboratory. We utilized single-cell multi-omics sequencing technology at the Biomedical Pioneering Innovation Center, School of Life Sciences, Peking University, to analyse the blastocysts. Transcriptome sequencing was used to predict the CNV of each cell through bioinformatics analysis, and the results were validated using the DNA methylation library of each cell to confirm chromosomal normalcy. We conducted statistical analysis and graphical plotting using R 4.2.1, SPSS 27, and GraphPad Prism 9.3. Mean age of the volunteers, the blastocyst morphology, and the developmental ratewere similar in ST and ICSI groups. The blastocysts in the ST group had some additional chromosomal types that were prone to variations beyond those enriched in the blastocysts of the ICSI group. Finally, both Bland-Altman concordance test and kappa concordancetest showed good chromosomal concordance between TE and ICM in the ICSI blastocysts (kappa = 0.659, P < 0.05), but not in ST blastocysts (P = 1.000), suggesting that the TE in reconstituted embryos is not representative of ICM. Gene functional annotation (GO and KEGG analyses) suggests that there may be new or additional pathways for CNV generation in ST embryos compared to ICSI embryos. This study was mainly limited by the small sample size and the limitations of single-cell multi-omics sequencing technology. To select eligible single cells, some cells of the embryos were eliminated or not labelled, resulting in a loss of information about them. The findings of this study are innovative and exploratory. A larger sample size of human embryos (especially ST embryos) and more accurate molecular genetics techniques for detecting CNV in single cells are needed to validate our results. Our study justifies the routine clinical use of PGT-A in ICSI blastocysts, as we found that the TE is a good substitute for ICM in predicting chromosomal abnormalities. While PGT-A is not entirely accurate, our data demonstrate good clinical feasibility. This trial was able to provide correct genetic counselling to patients regarding the reliability of PGT-A. Regarding ST blastocysts, the increased mosaicism rate and the inability of the TE to represent the chromosomal copy number of the ICM are both biological characteristics that differentiate them from ICSI blastocysts. Currently, ST is not used clinically on a large scale to produce blastocysts. However, if ST becomes more widely used in the future, our study will be the first to demonstrate that the use of PGT-A in ST blastocysts may not be as accurate as PGT-A for ICSI blastocysts. This study was supported by grants from the National Key R&D Program of China (2018YFA0107601) and the National Key R&D Program of China (2018YFC1003003). The authors declare no conflict of interest. N/A.

Single-cell multi-omics sequencing of human spermatogenesis reveals a DNA demethylation event associated with male meiotic recombination.

Human spermatogenesis is a highly ordered process; however, the roles of DNA methylation and chromatin accessibility in this process remain largely unknown. Here by simultaneously investigating the chromatin accessibility, DNA methylome and transcriptome landscapes using the modified single-cell chromatin overall omic-scale landscape sequencing approach, we revealed that the transcriptional changes throughout human spermatogenesis were correlated with chromatin accessibility changes. In particular, we identified a set of transcription factors and cis elements with potential functions. A round of DNA demethylation was uncovered upon meiosis initiation in human spermatogenesis, which was associated with male meiotic recombination and conserved between human and mouse. Aberrant DNA hypermethylation could be detected in leptotene spermatocytes of certain nonobstructive azoospermia patients. Functionally, the intervention of DNA demethylation affected male meiotic recombination and fertility. Our work provides multi-omics landscapes of human spermatogenesis at single-cell resolution and offers insights into the association between DNA demethylation and male meiotic recombination.

Single-cell multi-omics topic embedding reveals cell-type-specific and COVID-19 severity-related immune signatures

The advent of single-cell multi-omics sequencing technology makes it possible for researchers to leverage multiple modalities for individual cells and explore cell heterogeneity. However, the high-dimensional, discrete, and sparse nature of the data make the downstream analysis particularly challenging. Here, we propose an interpretable deep learning method called moETM to perform integrative analysis of high-dimensional single-cell multimodal data. moETM integrates multiple omics data via a product-of-experts in the encoder and employs multiple linear decoders to learn the multi-omics signatures. moETM demonstrates superior performance compared with six state-of-the-art methods on seven publicly available datasets. By applying moETM to the scRNA+ scATAC data, we identified sequence motifs corresponding to the transcription factors regulating immune gene signatures. Applying moETM to CITE-seq data from the COVID-19 patients revealed not only known immune cell-type-specific signatures but also composite multi-omics biomarkers of critical conditions due to COVID-19, thus providing insights from both biological and clinical perspectives.

Single-cell third-generation sequencing-based multi-omics uncovers gene expression changes governed by ecDNA and structural variants in cancer cells.

Cancer cells often exhibit large-scale genomic variations, such as circular extrachromosomal DNA (ecDNA) and structural variants (SVs), which have been highly correlated with the initiation and progression of cancer. Currently, no adequate method exists to unveil how these variations regulate gene expression in heterogeneous cancer cell populations at a single-cell resolution. Here, we developed a single-cell multi-omics sequencing method, scGTP-seq, to analyse ecDNA and SVs using long-read sequencing technologies. We demonstrated that our method can efficiently detect ecDNA and SVs and illustrated how these variations affect transcriptomic changes in various cell lines. Finally, we applied and validated this method in a clinical sample of hepatocellular carcinoma (HCC), demonstrating a feasible way to monitor the evolution of ecDNA and SVs during cancer progression.

Single-cell multi-omics sequencing reveals the immunological disturbance underlying STAT3-V637M Hyper-IgE syndrome

Hyper-IgE syndrome (HIES) is a primary immunodeficiency characterized by, among others, the excessive production of IgE and repetitive bacterial/fungal infections. Mutations in STAT3, a transcription factor that orchestrates immune responses, may cause HIES, but the underlying mechanisms are not fully understood. Here, we used multi-omic approaches to comprehensively decipher the immune disturbance in a male HIES patient harboring STAT3-V637M. In his peripheral blood mononuclear cell (PBMC) we found significant clonal expansion of CD8 T cells (with increased CD8 subunits expression, potentially enhancing responsiveness to MHC I molecules), but not in his CD4 T cells and B cells. Although his B cells exhibited a higher potential in producing immunoglobulin, elevated SPIC binding might bias the products toward IgE isotype. Immune checkpoint inhibitors, including CTLA4, LAG3, were overexpressed in his PBMC-CD4 T cells, accompanied by reduced CD28 and IL6ST (gp130) expression. In his CD4 T cells, integrative analyses predicted upstream transcription factors (including ETV6, KLF13, and RORA) for LAG3, IL6ST, and CD28, respectively. The down-regulation of phagocytosis and nitric oxide synthesis-related genes in his PBMC-monocytes seem to be the culprit of his disseminated bacterial/fungal infection. Counterintuitively, in his PBMC we predicted increased STAT3 binding in both naïve and mature CD4 compartments, although this was not observed in most of his PBMC. In his bronchoalveolar lavage fluid (BALF), we found two macrophage subtypes with anti-bacterial properties, which were identified by CXCL8/S100A8/S100A9, or SOD2, respectively. Together, we described how the immune cell landscape was disturbed in STAT3-V637M HIES, providing a resource for further studies.

Integrative single-cell multiomics analyses dissect molecular signatures of intratumoral heterogeneities and differentiation states of human gastric cancer.

Human gastric cancer is a highly lethal disease, but the underlying multiomic molecular signatures remain largely unclear. Here, we performed multi-regional sampling, parallel single-cell multiomics sequencing and integrated analyses of human gastric cancer. We identified common transcriptomic alterations of gastric cancer cells, such as aberrant down-regulation of genes associated with normal stomach function and up-regulation of KRT7, PI3, S100A4, etc. Surprisingly, aberrant and prevalent up-regulation of genes highly expressed in normal colorectal epithelial cells were also identified in cancer cells, which may be partially regulated by promoter chromatin accessibility and DNA methylation levels. We revealed the single-cell DNA methylome landscape of gastric cancer, and identified candidate DNA methylation biomarkers, such as hypermethylated promoters of TMEM240 and HAGLROS, and hypomethylated promoters of TRPM2-AS and HRH1. Additionally, the relationships between genetic lineages, DNA methylation and transcriptomic clusters were systematically revealed at single-cell level. We showed that DNA methylation heterogeneities were mainly among different genetic lineages of cancer cells. Moreover, we found that DNA methylation levels of cancer cells with poorer differentiation states tend to be higher than those of cancer cells with better differentiation states in the primary tumor within the same patient, although still lower than in normal gastric epithelial cells. Cancer cells with poorer differentiation states also prevalently down-regulated MUC1 expression and immune-related pathways, and had poor infiltration of CD8+ T cells. Our study dissected the molecular signatures of intratumoral heterogeneities and differentiation states of human gastric cancer using integrative single-cell multiomics analyses.

Dissecting the molecular features of bovine-arrested eight-cell embryos using single-cell multi-omics sequencing†.

The regulation of mammalian early-embryonic development is a complex, coordinated process that involves widespread transcriptomic and epigenetic remodeling. The main cause of developmental failure in preimplantation embryos after in vitro fertilization is the irreversible arrested-at-cleavage stage. To deepen our understanding of this embryonic block, we profiled a single-cell multi-omics map of copy number variations (CNVs), the transcriptome, the DNA methylome, and the chromatin state of bovine eight-cell embryos with a two-cell fate that either arrested or developed into blastocysts. To do this, we sequenced a biopsied blastomere and tracked the developmental potential of the remaining cells. Aneuploid embryos inferred by CNVs from DNA- and RNA-library data tended to lose their developmental potency. Analysis of distinct genomic regions of DNA methylation and chromatin accessibility revealed that enrichment of gene function and signaling pathways, such as the MAPK signaling pathway, was altered in arrested euploid eight-cell embryos compared with blastocyst-developed euploid eight-cell embryos. Moreover, the RNA expression and chromatin accessibility of embryonic genome activation-associated genes were lower in arrested euploid embryos than in blastocyst-developed embryos. Taken together, our results indicate that the developmental block of eight-cell embryos can be caused by multiple molecular layers, including CNVs, abnormality of DNA methylation and chromatin accessibility, and insufficient expression of embryonic genome activation-associated genes. Our integrated and comprehensive data set provides a valuable resource to further dissect the exact mechanisms underlying the arrest of bovine eight-cell embryos in vitro.

Enzalutamide-induced signatures revealed by epigenetic plasticity using single-cell multi-omics sequencing in prostate cancer

Prostate cancer is morphologically and molecularly heterogeneous, which poses obstacles for early diagnosis and treatment. Advancements in understanding the heterogeneity of prostate cancer will help navigate through these challenges and ultimately benefit patients. In this study, we integrated single-cell sequencing for transposase-accessible chromatin and whole transcriptome in prostate cancer cell lines, aiming to decode the epigenetic plasticity upon enzalutamide (ENZ) treatment. By comparing the cell populations representing early-treatment response or resistance to the initial tumor cells, we identified seven signature gene sets; they present consistent trends of chromatin closing co-occurred with down-regulated genes during early response and chromatin opening with up-regulated genes upon maintaining drug resistance. In the molecular signatures, we found genes ZNF337, MAPK15, and ESRRG are favorable in progression-free prognosis during early response, while genes CCDC150, CCDC18, and POC1A marked poor prognosis underpinning the pre-existing drug resistance in The Cancer Genome Atlas prostate adenocarcinoma cohort. Ultimately, drug-target analyses nominated combinatory drug candidates to either enhance early-treatment response or potentially overcome ENZ resistance. Together, our integrative, single-cell multi-omics approach in pre-clinical models is effective in identifying informative signatures from complex molecular events, illustrating diverse drug responses in prostate cancer, and invoking novel combinatory drug strategies to inform clinical decision making.

Single-cell multi-omics integration for unpaired data by a siamese network with graph-based contrastive loss

BackgroundSingle-cell omics technology is rapidly developing to measure the epigenome, genome, and transcriptome across a range of cell types. However, it is still challenging to integrate omics data from different modalities. Here, we propose a variation of the Siamese neural network framework called MinNet, which is trained to integrate multi-omics data on the single-cell resolution by using graph-based contrastive loss.ResultsBy training the model and testing it on several benchmark datasets, we showed its accuracy and generalizability in integrating scRNA-seq with scATAC-seq, and scRNA-seq with epitope data. Further evaluation demonstrated our model's unique ability to remove the batch effect, a common problem in actual practice. To show how the integration impacts downstream analysis, we established model-based smoothing and cis-regulatory element-inferring method and validated it with external pcHi-C evidence. Finally, we applied the framework to a COVID-19 dataset to bolster the original work with integration-based analysis, showing its necessity in single-cell multi-omics research.ConclusionsMinNet is a novel deep-learning framework for single-cell multi-omics sequencing data integration. It ranked top among other methods in benchmarking and is especially suitable for integrating datasets with batch and biological variances. With the single-cell resolution integration results, analysis of the interplay between genome and transcriptome can be done to help researchers understand their data and question.