Unique regulatory network of dragon fruit simultaneously mitigates the effect of vanadium pollutant and environmental factors.

Under changing climatic conditions, plants are simultaneously facing conflicting stresses in nature. Plants can sense different stresses, induce systematic ROS signals, and regulate transcriptomic, hormonal, and stomatal responses. We performed transcriptome analysis to reveal the integrative stress response regulatory mechanism underlying heavy metal stress alone or in combination with heat and drought conditions in pitaya (dragon fruit). A total of 70 genes were identified from 31,130 transcripts with conserved differential expression. Furthermore, weighted gene co-expression network analysis (WGCNA) identified trait-associated modules. By integrating information from three modules and protein-protein interaction (PPI) networks, we identified 10 interconnected genes associated with the multifaceted defense mechanism employed by pitaya against co-occurring stresses. To further confirm the reliability of the results, we performed a comparative analysis of 350 genes identified by three trait modules and 70 conserved genes exhibiting their dynamic expression under all treatments. Differential expression pattern of genes and comparative analysis, have proven instrumental in identifying ten putative structural genes. These ten genes were annotated as PLAT/LH2, CAT, MLP, HSP, PB1, PLA, NAC, HMA, and CER1 transcription factors involved in antioxidant activity, defense response, MAPK signaling, detoxification of metals and regulating the crosstalk between the complex pathways. Predictive analysis of putative candidate genes, potentially governing single, double, and multifactorial stress response, by several signaling systems and molecular patterns. These findings represent a valuable resource for pitaya breeding programs, offering the potential to develop resilient "super pitaya" plants.

Leveraging a neutrophil-derived PCD signature to predict and stratify patients with acute myocardial infarction: from AI prediction to biological interpretation.

BACKGROUND: Programmed cell death (PCD) has recently been implicated in modulating the removal of neutrophils recruited in acute myocardial infarction (AMI). Nonetheless, the clinical significance and biological mechanism of neutrophil-related PCD remain unexplored. METHODS: We employed an integrative machine learning-based computational framework to generate a predictive neutrophil-derived PCD signature (NPCDS) within five independent microarray cohorts from the peripheral blood of AMI patients. Non-negative matrix factorization was leveraged to develop an NPCDS-based AMI subtype. To elucidate the biological mechanism underlying NPCDS, we implemented single-cell transcriptomics on Cd45+ cells isolated from the murine heart of experimental AMI. We finally conducted a Mendelian randomization (MR) study and molecular docking to investigate the therapeutic value of NPCDS on AMI. RESULTS: We reported the robust and superior performance of NPCDS in AMI prediction, which contributed to an optimal combination of random forest and stepwise regression fitted on nine neutrophil-related PCD genes (MDM2, PTK2B, MYH9, IVNS1ABP, MAPK14, GNS, MYD88, TLR2, CFLAR). Two divergent NPCDS-based subtypes of AMI were revealed, in which subtype 1 was characterized as inflammation-activated with more vibrant neutrophil activities, whereas subtype 2 demonstrated the opposite. Mechanically, we unveiled the expression dynamics of NPCDS to regulate neutrophil transformation from a pro-inflammatory phase to an anti-inflammatory phase in AMI. We uncovered a significant causal association between genetic predisposition towards MDM2 expression and the risk of AMI. We also found that lidoflazine, isotetrandrine, and cepharanthine could stably target MDM2. CONCLUSION: Altogether, NPCDS offers significant implications for prediction, stratification, and therapeutic management for AMI.

The whole transcriptome analysis using FFPE and fresh tissue samples identifies the molecular fingerprint of osteosarcoma.

Osteosarcoma is a form of bone cancer that predominantly impacts osteoblasts, the cells responsible for creating fresh bone tissue. Typical indications include bone pain, inflammation, sensitivity, mobility constraints, and fractures. Utilising imaging techniques such as X-rays, MRI scans, and CT scans can provide insights into the size and location of the tumour. Additionally, a biopsy is employed to confirm the diagnosis. Analysing genes with distinct expression patterns unique to osteosarcoma can be valuable for early detection and the development of effective treatment approaches. In this research, we comprehensively examined the entire transcriptome and pinpointed genes with altered expression profiles specific to osteosarcoma. The study mainly aimed to identify the molecular fingerprint of osteosarcoma. In this study, we processed 90 FFPE samples from PathWest with an almost equal number of osteosarcoma and healthy tissues. RNA was extracted from Paraffin-embedded tissue; RNA was sequenced, the sequencing data was analysed, and gene expression was compared to the healthy samples of the same patients. Differentially expressed genes in osteosarcoma-derived samples were identified, and the functions of those genes were explored. This result was combined with our previous studies based on FFPE and fresh samples to perform a meta-analysis. We identified 1,500 identical differentially expressed genes in PathWest osteosarcoma samples compared to normal tissue samples of the same patients. Meta-analysis with combined fresh tissue samples identified 530 differentially expressed genes. IFITM5, MMP13, PANX3, and MAGEA6 were some of the most overexpressed genes in osteosarcoma samples, while SLC4A1, HBA1, HBB, AQP7 genes were some of the top downregulated genes. Through the meta-analysis, 530 differentially expressed genes were identified to be identical among FFPE (105 FFPE samples) and 36 fresh bone samples. Deconvolution analysis with single-cell RNAseq data confirmed the presence of specific cell clusters in FFPE samples. We propose these 530 DEGs as a molecular fingerprint of osteosarcoma.

The chromatin accessibility and transcriptomic landscape of the aging mice cochlea and the identification of potential functional super-enhancers in age-related hearing loss.

BACKGROUND: Presbycusis, also referred to as age-related hearing loss (ARHL), is a condition that results from the cumulative effects of aging on an individual's auditory capabilities. Given the limited understanding of epigenetic mechanisms in ARHL, our research focuses on alterations in chromatin-accessible regions. METHODS: We employed assay for transposase-accessible chromatin with high-throughput sequencing (ATAC-seq) in conjunction with unique identifier (UID) mRNA-seq between young and aging cochleae, and conducted integrated analysis as well as motif/TF-gene prediction. Additionally, the essential role of super-enhancers (SEs) in the development of ARHL was identified by comparative analysis to previous research. Meanwhile, an ARHL mouse model and an aging mimic hair cell (HC) model were established with a comprehensive identification of senescence phenotypes to access the role of SEs in ARHL progression. RESULTS: The control cochlear tissue exhibited greater chromatin accessibility than cochlear tissue affected by ARHL. Furthermore, the levels of histone 3 lysine 27 acetylation were significantly depressed in both aging cochlea and aging mimic HEI-OC1 cells, highlighting the essential role of SEs in the development of ARHL. The potential senescence-associated super-enhancers (SASEs) of ARHL were identified, most of which exhibited decreased chromatin accessibility. The majority of genes related to the SASEs showed obvious decreases in mRNA expression level in aging HCs and was noticeably altered following treatment with JQ1 (a commonly used SE inhibitor). CONCLUSION: The chromatin accessibility in control cochlear tissue was higher than that in cochlear tissue affected by ARHL. Potential SEs involved in ARHL were identified, which might provide a basis for future therapeutics targeting SASEs related to ARHL.

Identification of core genes shared by ischemic stroke and myocardial infarction using an integrated approach.

BACKGROUND: Both ischemic stroke (IS) and myocardial infarction (MI) are caused by vascular occlusion that results in ischemia. While there may be similarities in their mechanisms, the potential relationship between these 2 diseases has not been comprehensively analyzed. Therefore, this study explored the commonalities in the pathogenesis of IS and MI. METHODS: Datasets for IS (GSE58294, GSE16561) and MI (GSE60993, GSE61144) were downloaded from the Gene Expression Omnibus database. Transcriptome data from each of the 4 datasets were analyzed using bioinformatics, and the differentially expressed genes (DEGs) shared between IS and MI were identified and subsequently visualized using a Venn diagram. A protein-protein interaction (PPI) network was constructed using the Interacting Gene Retrieval Tool database, and identification of key core genes was performed using CytoHubba. Gene Ontology (GO) term annotation and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis of the shared DEGs were conducted using prediction and network analysis methods, and the functions of the hub genes were determined using Metascape. RESULTS: The analysis revealed 116 and 1321 DEGs in the IS and MI datasets, respectively. Of the 75 DEGs shared between IS and MI, 56 were upregulated and 19 were downregulated. Furthermore, 15 core genes - S100a12, Hp, Clec4d, Cd163, Mmp9, Ormdl3, Il2rb, Orm1, Irak3, Tlr5, Lrg1, Clec4e, Clec5a, Mcemp1, and Ly96 - were identified. GO enrichment analysis of the DEGs showed that they were mainly involved in the biological functions of neutrophil degranulation, neutrophil activation during immune response, and cytokine secretion. KEGG analysis showed enrichment in pathways pertaining to Salmonella infection, Legionellosis, and inflammatory bowel disease. Finally, the core gene-transcription factor, gene-microRNA, and small-molecule relationships were predicted. CONCLUSION: These core genes may provide a novel theoretical basis for the diagnosis and treatment of IS and MI.

Transcriptomic Analysis During the Induction of Somatic Embryogenesis in Coffea canephora.

Omic tools have changed the way of doing research in experimental biology. The somatic embryogenesis (SE) study has not been immune to this benefit. The transcriptomic tools have been used to compare the genes expressed during the induction of SE with the genes expressed in zygotic embryogenesis or to compare the development of the different stages embryos go through. It has also been used to compare the expression of genes during the development of calli from which SE is induced, as well as many other applications. The protocol described here is employed in our laboratory to extract RNA and generate several transcriptomes for the study of SE on Coffea canephora.

A strategy to detect metabolic changes induced by exposure to chemicals from large sets of condition-specific metabolic models computed with enumeration techniques.

BACKGROUND: The growing abundance of in vitro omics data, coupled with the necessity to reduce animal testing in the safety assessment of chemical compounds and even eliminate it in the evaluation of cosmetics, highlights the need for adequate computational methodologies. Data from omics technologies allow the exploration of a wide range of biological processes, therefore providing a better understanding of mechanisms of action (MoA) related to chemical exposure in biological systems. However, the analysis of these large datasets remains difficult due to the complexity of modulations spanning multiple biological processes. RESULTS: To address this, we propose a strategy to reduce information overload by computing, based on transcriptomics data, a comprehensive metabolic sub-network reflecting the metabolic impact of a chemical. The proposed strategy integrates transcriptomic data to a genome scale metabolic network through enumeration of condition-specific metabolic models hence translating transcriptomics data into reaction activity probabilities. Based on these results, a graph algorithm is applied to retrieve user readable sub-networks reflecting the possible metabolic MoA (mMoA) of chemicals. This strategy has been implemented as a three-step workflow. The first step consists in building cell condition-specific models reflecting the metabolic impact of each exposure condition while taking into account the diversity of possible optimal solutions with a partial enumeration algorithm. In a second step, we address the challenge of analyzing thousands of enumerated condition-specific networks by computing differentially activated reactions (DARs) between the two sets of enumerated possible condition-specific models. Finally, in the third step, DARs are grouped into clusters of functionally interconnected metabolic reactions, representing possible mMoA, using the distance-based clustering and subnetwork extraction method. The first part of the workflow was exemplified on eight molecules selected for their known human hepatotoxic outcomes associated with specific MoAs well described in the literature and for which we retrieved primary human hepatocytes transcriptomic data in Open TG-GATEs. Then, we further applied this strategy to more precisely model and visualize associated mMoA for two of these eight molecules (amiodarone and valproic acid). The approach proved to go beyond gene-based analysis by identifying mMoA when few genes are significantly differentially expressed (2 differentially expressed genes (DEGs) for amiodarone), bringing additional information from the network topology, or when very large number of genes were differentially expressed (5709 DEGs for valproic acid). In both cases, the results of our strategy well fitted evidence from the literature regarding known MoA. Beyond these confirmations, the workflow highlighted potential other unexplored mMoA. CONCLUSION: The proposed strategy allows toxicology experts to decipher which part of cellular metabolism is expected to be affected by the exposition to a given chemical. The approach originality resides in the combination of different metabolic modelling approaches (constraint based and graph modelling). The application to two model molecules shows the strong potential of the approach for interpretation and visual mining of complex omics in vitro data. The presented strategy is freely available as a python module ( https://pypi.org/project/manamodeller/ ) and jupyter notebooks ( https://github.com/LouisonF/MANA ).

Global Transcriptomic and Characteristics Comparisons between Mouse Fetal Liver and Bone Marrow Definitive Erythropoiesis.

Erythropoiesis occurs first in the yolk sac as a transit "primitive" form, then is gradually replaced by the "definitive" form in the fetal liver (FL) during fetal development and in the bone marrow (BM) postnatally. While it is well known that differences exist between primitive and definitive erythropoiesis, the similarities and differences between FL and BM definitive erythropoiesis have not been studied. Here we performed comprehensive comparisons of erythroid progenitors and precursors at all maturational stages sorted from E16.5 FL and adult BM. We found that FL cells at all maturational stages were larger than their BM counterparts. We further found that FL BFU-E cells divided at a faster rate and underwent more cell divisions than BM BFU-E. Transcriptome comparison revealed that genes with increased expression in FL BFU-Es were enriched in cell division. Interestingly, the expression levels of glucocorticoid receptor Nr3c1, Myc and Myc downstream target Ccna2 were significantly higher in FL BFU-Es, indicating the role of the Nr3c1-Myc-Ccna2 axis in the enhanced proliferation/cell division of FL BFU-E cells. At the CFU-E stage, the expression of genes associated with hemoglobin biosynthesis were much higher in FL CFU-Es, indicating more hemoglobin production. During terminal erythropoiesis, overall temporal patterns in gene expression were conserved between the FL and BM. While biological processes related to translation, the tricarboxylic acid cycle and hypoxia response were upregulated in FL erythroblasts, those related to antiviral signal pathway were upregulated in BM erythroblasts. Our findings uncovered previously unrecognized differences between FL and BM definitive erythropoiesis and provide novel insights into erythropoiesis.

Opportunities and Challenges in Advancing Plant Research with Single-cell Omics.

Plants possess diverse cell types and intricate regulatory mechanisms to adapt to the ever-changing environment of nature. Various strategies have been employed to study cell types and their developmental progressions, including single-cell sequencing methods which provide high-dimensional catalogs to address biological concerns. In recent years, single-cell sequencing technologies in transcriptomics, epigenomics, proteomics, metabolomics, and spatial transcriptomics have been increasingly used in plant science to reveal intricate biological relationships at the single-cell level. However, the application of single-cell technologies to plants is more limited due to the challenges posed by cell structure. This review outlines the advancements in single-cell omics technologies, their implications in plant systems, future research applications, and the challenges of single-cell omics in plant systems.

The myelin water imaging transcriptome: myelin water fraction regionally varies with oligodendrocyte-specific gene expression.

Identifying sensitive and specific measures that can quantify myelin are instrumental in characterizing microstructural changes in neurological conditions. Neuroimaging transcriptomics is emerging as a valuable technique in this regard, offering insights into the molecular basis of promising candidates for myelin quantification, such as myelin water fraction (MWF). We aimed to demonstrate the utility of neuroimaging transcriptomics by validating MWF as a myelin measure. We utilized data from a normative MWF brain atlas, comprised of 50 healthy subjects (mean age = 25 years, range = 17-42 years) scanned at 3 Tesla. Magnetic resonance imaging data included myelin water imaging to extract MWF and T1 anatomical scans for image registration and segmentation. We investigated the inter-regional distributions of gene expression data from the Allen Human Brain Atlas in conjunction with inter-regional MWF distribution patterns. Pearson correlations were used to identify genes with expression profiles mirroring MWF. The Single Cell Type Atlas from the Human Protein Atlas was leveraged to classify genes into gene sets with high cell type specificity, and a control gene set with low cell type specificity. Then, we compared the Pearson correlation coefficients for each gene set to determine if cell type-specific gene expression signatures correlate with MWF. Pearson correlation coefficients between MWF and gene expression for oligodendrocytes and adipocytes were significantly higher than for the control gene set, whereas correlations between MWF and inhibitory/excitatory neurons were significantly lower. Our approach in integrating transcriptomics with neuroimaging measures supports an emerging technique for understanding and validating MRI-derived markers such as MWF.

Trehalase inhibition in Helicoverpa armigera activates machinery for alternate energy acquisition.

Trehalose serves as a primary circulatory sugar in insects which is crucial in energy metabolism and stress recovery. It is hydrolyzed into two glucose molecules by trehalase. Silencing or inhibiting trehalase results in reduced fitness, developmental defects, and insect mortality. Despite its importance, the molecular response of insects to trehalase inhibition is not known. Here, we performed transcriptomic analyses of Helicoverpa armigera treated with validamycin A (VA), a trehalase inhibitor. VA ingestion resulted in increased mortality, developmental delay, and reduced ex vivo trehalase activity. Pathway enrichment and gene ontology analyses suggest that key genes involved in carbohydrate, protein, fatty acid, and mitochondria-related metabolisms are deregulated. The activation of protein and fat degradation may be necessary to fulfil energy requirements, evidenced by the dysregulated expression of critical genes in these metabolisms. Co-expression analysis supports the notion that trehalase inhibition leads to putative interaction with key regulators of other pathways. Metabolomics correlates with transcriptomics to show reduced levels of key energy metabolites. VA generates an energy-deficient condition, and insects activate alternate pathways to facilitate the energy demand. Overall, this study provides insights into the molecular mechanisms underlying the response of insects to trehalase inhibition and highlights potential targets for insect control.

LSTrAP-denovo: Automated Generation of Transcriptome Atlases for Eukaryotic Species Without Genomes.

Despite the abundance of species with transcriptomic data, a significant number of species still lack sequenced genomes, making it difficult to study gene function and expression in these organisms. While de novo transcriptome assembly can be used to assemble protein-coding transcripts from RNA-sequencing (RNA-seq) data, the datasets used often only feature samples of arbitrarily selected or similar experimental conditions, which might fail to capture condition-specific transcripts. We developed the Large-Scale Transcriptome Assembly Pipeline for de novo assembled transcripts (LSTrAP-denovo) to automatically generate transcriptome atlases of eukaryotic species. Specifically, given an NCBI TaxID, LSTrAP-denovo can (1) filter undesirable RNA-seq accessions based on read data, (2) select RNA-seq accessions via unsupervised machine learning to construct a sample-balanced dataset for download, (3) assemble transcripts via over-assembly, (4) functionally annotate coding sequences (CDS) from assembled transcripts and (5) generate transcriptome atlases in the form of expression matrices for downstream transcriptomic analyses. LSTrAP-denovo is easy to implement, written in Python, and is freely available at https://github.com/pengkenlim/LSTrAP-denovo/.

Integrative metagenomic, transcriptomic, and proteomic analysis reveal the microbiota-host interplay in early-stage lung adenocarcinoma among non-smokers.

BACKGROUND: The incidence of early-stage lung adenocarcinoma (ES-LUAD) is steadily increasing among non-smokers. Previous research has identified dysbiosis in the gut microbiota of patients with lung cancer. However, the local microbial profile of non-smokers with ES-LUAD remains largely unknown. In this study, we systematically characterized the local microbial community and its associated features to enable early intervention. METHODS: A prospective collection of ES-LUAD samples (46 cases) and their corresponding normal tissues adjacent to the tumor (41 cases), along with normal lung tissue samples adjacent to pulmonary bullae in patients with spontaneous pneumothorax (42 cases), were subjected to ultra-deep metagenomic sequencing, host transcriptomic sequencing, and proteomic sequencing. The obtained omics data were subjected to both individual and integrated analysis using Spearman correlation coefficients. RESULTS: We concurrently detected the presence of bacteria, fungi, and viruses in the lung tissues. The microbial profile of ES-LUAD exhibited similarities to NAT but demonstrated significant differences from the healthy controls (HCs), characterized by an overall reduction in species diversity. Patients with ES-LUAD exhibited local microbial dysbiosis, suggesting the potential pathogenicity of certain microbial species. Through multi-omics correlations, intricate local crosstalk between the host and local microbial communities was observed. Additionally, we identified a significant positive correlation (rho > 0.6) between Methyloversatilis discipulorum and GOLM1 at both the transcriptional and protein levels using multi-omics data. This correlated axis may be associated with prognosis. Finally, a diagnostic model composed of six bacterial markers successfully achieved precise differentiation between patients with ES-LUAD and HCs. CONCLUSIONS: Our study depicts the microbial spectrum in patients with ES-LUAD and provides evidence of alterations in lung microbiota and their interplay with the host, enhancing comprehension of the pathogenic mechanisms that underlie ES-LUAD. The specific model incorporating lung microbiota can serve as a potential diagnostic tool for distinguishing between ES-LUAD and HCs.

Comparative Transcriptome Analysis Reveals Key Functions of MiMYB Gene Family in Macadamia Nut Pericarp Formation.

Macadamia nuts are one of the most important economic food items in the world. Pericarp thickness and flavonoid composition are the key quality traits of Macadamia nuts, but the underlying mechanism of pericarp formation is still unknown. In this study, three varieties with significantly different pericarp thicknesses, namely, A38, Guire No.1, and HAES 900, at the same stage of maturity, were used for transcriptome analysis, and the results showed that there were significant differences in their gene expression profile. A total of 3837 new genes were discovered, of which 1532 were functionally annotated. The GO, COG, and KEGG analysis showed that the main categories in which there were significant differences were flavonoid biosynthesis, phenylpropanoid biosynthesis, and the cutin, suberine, and wax biosynthesis pathways. Furthermore, 63 MiMYB transcription factors were identified, and 56 R2R3-MYB transcription factors were clustered into different subgroups compared with those in Arabidopsis R2R3-MYB. Among them, the S4, S6, and S7 subgroups were involved in flavonoid biosynthesis and pericarp formation. A total of 14 MiMYBs' gene expression were verified by RT-qPCR analysis. These results provide fundamental knowledge of the pericarp formation regulatory mechanism in macadamia nuts.

Transcriptome Dynamics and Cell Dialogs Between Oocytes and Granulosa Cells in Mouse Follicle Development.

The development and maturation of follicles is a sophisticated and multistage process. The dynamic gene expression of oocytes and their surrounding somatic cells and the dialogs between these cells are critical to this process. In this study, we accurately classified the oocyte and follicle development into nine stages and profiled the gene expression of mouse oocytes and their surrounding granulosa cells and cumulus cells. The clustering of the transcriptomes showed the trajectories of two distinct development courses of oocytes and their surrounding somatic cells. Gene expression changes precipitously increased at Type 4 stage and drastically dropped afterward within both oocytes and granulosa cells. Moreover, the number of differentially expressed genes between oocytes and granulosa cells dramatically increased at Type 4 stage, most of which persistently passed on to the later stages. Strikingly, cell communications within and between oocytes and granulosa cells became active from Type 4 stage onward. Cell dialogs connected oocytes and granulosa cells in both unidirectional and bidirectional manners. TGFB2/3, TGFBR2/3, INHBA/B, and ACVR1/1B/2B of TGF-ß signaling pathway functioned in the follicle development. NOTCH signaling pathway regulated the development of granulosa cells. Additionally, many maternally DNA methylation- or H3K27me3-imprinted genes remained active in granulosa cells but silent in oocytes during oogenesis. Collectively, Type 4 stage is the key turning point when significant transcription changes diverge the fate of oocytes and granulosa cells, and the cell dialogs become active to assure follicle development. These findings shed new insights on the transcriptome dynamics and cell dialogs facilitating the development and maturation of oocytes and follicles.

DeepGSEA: explainable deep gene set enrichment analysis for single-cell transcriptomic data.

MOTIVATION: Gene set enrichment (GSE) analysis allows for an interpretation of gene expression through pre-defined gene set databases and is a critical step in understanding different phenotypes. With the rapid development of single-cell RNA sequencing (scRNA-seq) technology, GSE analysis can be performed on fine-grained gene expression data to gain a nuanced understanding of phenotypes of interest. However, with the cellular heterogeneity in single-cell gene profiles, current statistical GSE analysis methods sometimes fail to identify enriched gene sets. Meanwhile, deep learning has gained traction in applications like clustering and trajectory inference in single-cell studies due to its prowess in capturing complex data patterns. However, its use in GSE analysis remains limited, due to interpretability challenges. RESULTS: In this paper, we present DeepGSEA, an explainable deep gene set enrichment analysis approach which leverages the expressiveness of interpretable, prototype-based neural networks to provide an in-depth analysis of GSE. DeepGSEA learns the ability to capture GSE information through our designed classification tasks, and significance tests can be performed on each gene set, enabling the identification of enriched sets. The underlying distribution of a gene set learned by DeepGSEA can be explicitly visualized using the encoded cell and cellular prototype embeddings. We demonstrate the performance of DeepGSEA over commonly used GSE analysis methods by examining their sensitivity and specificity with four simulation studies. In addition, we test our model on three real scRNA-seq datasets and illustrate the interpretability of DeepGSEA by showing how its results can be explained. AVAILABILITY AND IMPLEMENTATION: https://github.com/Teddy-XiongGZ/DeepGSEA.

SFINN: inferring gene regulatory network from single-cell and spatial transcriptomic data with shared factor neighborhood and integrated neural network.

MOTIVATION: The rise of single-cell RNA sequencing (scRNA-seq) technology presents new opportunities for constructing detailed cell type-specific gene regulatory networks (GRNs) to study cell heterogeneity. However, challenges caused by noises, technical errors, and dropout phenomena in scRNA-seq data pose significant obstacles to GRN inference, making the design of accurate GRN inference algorithms still essential. The recent growth of both single-cell and spatial transcriptomic sequencing data enables the development of supervised deep learning methods to infer GRNs on these diverse single-cell datasets. RESULTS: In this study, we introduce a novel deep learning framework based on shared factor neighborhood and integrated neural network (SFINN) for inferring potential interactions and causalities between transcription factors and target genes from single-cell and spatial transcriptomic data. SFINN utilizes shared factor neighborhood to construct cellular neighborhood network based on gene expression data and additionally integrates cellular network generated from spatial location information. Subsequently, the cell adjacency matrix and gene pair expression are fed into an integrated neural network framework consisting of a graph convolutional neural network and a fully-connected neural network to determine whether the genes interact. Performance evaluation in the tasks of gene interaction and causality prediction against the existing GRN reconstruction algorithms demonstrates the usability and competitiveness of SFINN across different kinds of data. SFINN can be applied to infer GRNs from conventional single-cell sequencing data and spatial transcriptomic data. AVAILABILITY AND IMPLEMENTATION: SFINN can be accessed at GitHub: https://github.com/JGuan-lab/SFINN.

A variational autoencoder trained with priors from canonical pathways increases the interpretability of transcriptome data.

Interpreting transcriptome data is an important yet challenging aspect of bioinformatic analysis. While gene set enrichment analysis is a standard tool for interpreting regulatory changes, we utilize deep learning techniques, specifically autoencoder architectures, to learn latent variables that drive transcriptome signals. We investigate whether simple, variational autoencoder (VAE), and beta-weighted VAE are capable of learning reduced representations of transcriptomes that retain critical biological information. We propose a novel VAE that utilizes priors from biological data to direct the network to learn a representation of the transcriptome that is based on understandable biological concepts. After benchmarking five different autoencoder architectures, we found that each succeeded in reducing the transcriptomes to 50 latent dimensions, which captured enough variation for accurate reconstruction. The simple, fully connected autoencoder, performs best across the benchmarks, but lacks the characteristic of having directly interpretable latent dimensions. The beta-weighted, prior-informed VAE implementation is able to solve the benchmarking tasks, and provide semantically accurate latent features equating to biological pathways. This study opens a new direction for differential pathway analysis in transcriptomics with increased transparency and interpretability.

Long-read RNA-Seq for the discovery of long noncoding and antisense RNAs in plant organelles.

Plant organelle transcription has been studied for decades. As techniques advanced, so did the fields of mitochondrial and plastid transcriptomics. The current view is that organelle genomes are pervasively transcribed, irrespective of their size, content, structure, and taxonomic origin. However, little is known about the nature of organelle noncoding transcriptomes, including pervasively transcribed noncoding RNAs (ncRNAs). Next-generation sequencing data have uncovered small ncRNAs in the organelles of plants and other organisms, but long ncRNAs remain poorly understood. Here, we argue that publicly available third-generation long-read RNA sequencing data from plants can provide a fine-tuned picture of long ncRNAs within organelles. Indeed, given their bloated architectures, plant mitochondrial genomes are well suited for studying pervasive transcription of ncRNAs. Ultimately, we hope to showcase this new avenue of plant research while also underlining the limitations of the proposed approach.