scMultiGAN: cell-specific imputation for single-cell transcriptomes with multiple deep generative adversarial networks.

The emergence of single-cell RNA sequencing (scRNA-seq) technology has revolutionized the identification of cell types and the study of cellular states at a single-cell level. Despite its significant potential, scRNA-seq data analysis is plagued by the issue of missing values. Many existing imputation methods rely on simplistic data distribution assumptions while ignoring the intrinsic gene expression distribution specific to cells. This work presents a novel deep-learning model, named scMultiGAN, for scRNA-seq imputation, which utilizes multiple collaborative generative adversarial networks (GAN). Unlike traditional GAN-based imputation methods that generate missing values based on random noises, scMultiGAN employs a two-stage training process and utilizes multiple GANs to achieve cell-specific imputation. Experimental results show the efficacy of scMultiGAN in imputation accuracy, cell clustering, differential gene expression analysis and trajectory analysis, significantly outperforming existing state-of-the-art techniques. Additionally, scMultiGAN is scalable to large scRNA-seq datasets and consistently performs well across sequencing platforms. The scMultiGAN code is freely available at https://github.com/Galaxy8172/scMultiGAN.

Machine learning meets omics: applications and perspectives.

The innovation of biotechnologies has allowed the accumulation of omics data at an alarming rate, thus introducing the era of 'big data'. Extracting inherent valuable knowledge from various omics data remains a daunting problem in bioinformatics. Better solutions often need some kind of more innovative methods for efficient handlings and effective results. Recent advancements in integrated analysis and computational modeling of multi-omics data helped address such needs in an increasingly harmonious manner. The development and application of machine learning have largely advanced our insights into biology and biomedicine and greatly promoted the development of therapeutic strategies, especially for precision medicine. Here, we propose a comprehensive survey and discussion on what happened, is happening and will happen when machine learning meets omics. Specifically, we describe how artificial intelligence can be applied to omics studies and review recent advancements at the interface between machine learning and the ever-widest range of omics including genomics, transcriptomics, proteomics, metabolomics, radiomics, as well as those at the single-cell resolution. We also discuss and provide a synthesis of ideas, new insights, current challenges and perspectives of machine learning in omics.

Detecting Allele-Specific Alternative Splicing from Population-Scale RNA-Seq Data.

RNA sequencing (RNA-seq) is a powerful technology for studying human transcriptome variation. We introduce PAIRADISE (Paired Replicate Analysis of Allelic Differential Splicing Events), a method for detecting allele-specific alternative splicing (ASAS) from RNA-seq data. Unlike conventional approaches that detect ASAS events one sample at a time, PAIRADISE aggregates ASAS signals across multiple individuals in a population. By treating the two alleles of an individual as paired, and multiple individuals sharing a heterozygous SNP as replicates, we formulate ASAS detection using PAIRADISE as a statistical problem for identifying differential alternative splicing from RNA-seq data with paired replicates. PAIRADISE outperforms alternative statistical models in simulation studies. Applying PAIRADISE to replicate RNA-seq data of a single individual and to population-scale RNA-seq data across many individuals, we detect ASAS events associated with genome-wide association study (GWAS) signals of complex traits or diseases. Additionally, PAIRADISE ASAS analysis detects the effects of rare variants on alternative splicing. PAIRADISE provides a useful computational tool for elucidating the genetic variation and phenotypic association of alternative splicing in populations.

scIGANs: single-cell RNA-seq imputation using generative adversarial networks.

Single-cell RNA-sequencing (scRNA-seq) enables the characterization of transcriptomic profiles at the single-cell resolution with increasingly high throughput. However, it suffers from many sources of technical noises, including insufficient mRNA molecules that lead to excess false zero values, termed dropouts. Computational approaches have been proposed to recover the biologically meaningful expression by borrowing information from similar cells in the observed dataset. However, these methods suffer from oversmoothing and removal of natural cell-to-cell stochasticity in gene expression. Here, we propose the generative adversarial networks (GANs) for scRNA-seq imputation (scIGANs), which uses generated cells rather than observed cells to avoid these limitations and balances the performance between major and rare cell populations. Evaluations based on a variety of simulated and real scRNA-seq datasets show that scIGANs is effective for dropout imputation and enhances various downstream analysis. ScIGANs is robust to small datasets that have very few genes with low expression and/or cell-to-cell variance. ScIGANs works equally well on datasets from different scRNA-seq protocols and is scalable to datasets with over 100 000 cells. We demonstrated in many ways with compelling evidence that scIGANs is not only an application of GANs in omics data but also represents a competing imputation method for the scRNA-seq data.

Deep learning of the splicing (epi)genetic code reveals a novel candidate mechanism linking histone modifications to ESC fate decision.

Alternative splicing (AS) is a genetically and epigenetically regulated pre-mRNA processing to increase transcriptome and proteome diversity. Comprehensively decoding these regulatory mechanisms holds promise in getting deeper insights into a variety of biological contexts involving in AS, such as development and diseases. We assembled splicing (epi)genetic code, DeepCode, for human embryonic stem cell (hESC) differentiation by integrating heterogeneous features of genomic sequences, 16 histone modifications with a multi-label deep neural network. With the advantages of epigenetic features, DeepCode significantly improves the performance in predicting the splicing patterns and their changes during hESC differentiation. Meanwhile, DeepCode reveals the superiority of epigenomic features and their dominant roles in decoding AS patterns, highlighting the necessity of including the epigenetic properties when assembling a more comprehensive splicing code. Moreover, DeepCode allows the robust predictions across cell lineages and datasets. Especially, we identified a putative H3K36me3-regulated AS event leading to a nonsense-mediated mRNA decay of BARD1. Reduced BARD1 expression results in the attenuation of ATM/ATR signalling activities and further the hESC differentiation. These results suggest a novel candidate mechanism linking histone modifications to hESC fate decision. In addition, when trained in different contexts, DeepCode can be expanded to a variety of biological and biomedical fields.

Identify bilayer modules via pseudo-3D clustering: applications to miRNA-gene bilayer networks.

Module identification is a frequently used approach for mining local structures with more significance in global networks. Recently, a wide variety of bilayer networks are emerging to characterize the more complex biological processes. In the light of special topological properties of bilayer networks and the accompanying challenges, there is yet no effective method aiming at bilayer module identification to probe the modular organizations from the more inspiring bilayer networks. To this end, we proposed the pseudo-3D clustering algorithm, which starts from extracting initial non-hierarchically organized modules and then iteratively deciphers the hierarchical organization of modules according to a bottom-up strategy. Specifically, a modularity function for bilayer modules was proposed to facilitate the algorithm reporting the optimal partition that gives the most accurate characterization of the bilayer network. Simulation studies demonstrated its robustness and outperformance against alternative competing methods. Specific applications to both the soybean and human miRNA-gene bilayer networks demonstrated that the pseudo-3D clustering algorithm successfully identified the overlapping, hierarchically organized and highly cohesive bilayer modules. The analyses on topology, functional and human disease enrichment and the bilayer subnetwork involved in soybean fat biosynthesis provided both the theoretical and biological evidence supporting the effectiveness and robustness of pseudo-3D clustering algorithm.

Inferring the soybean (Glycine max) microRNA functional network based on target gene network.

MOTIVATION: The rapid accumulation of microRNAs (miRNAs) and experimental evidence for miRNA interactions has ushered in a new area of miRNA research that focuses on network more than individual miRNA interaction, which provides a systematic view of the whole microRNome. So it is a challenge to infer miRNA functional interactions on a system-wide level and further draw a miRNA functional network (miRFN). A few studies have focused on the well-studied human species; however, these methods can neither be extended to other non-model organisms nor take fully into account the information embedded in miRNA-target and target-target interactions. Thus, it is important to develop appropriate methods for inferring the miRNA network of non-model species, such as soybean (Glycine max), without such extensive miRNA-phenotype associated data as miRNA-disease associations in human. RESULTS: Here we propose a new method to measure the functional similarity of miRNAs considering both the site accessibility and the interactive context of target genes in functional gene networks. We further construct the miRFNs of soybean, which is the first study on soybean miRNAs on the network level and the core methods can be easily extended to other species. We found that miRFNs of soybean exhibit a scale-free, small world and modular architecture, with their degrees fit best to power-law and exponential distribution. We also showed that miRNA with high degree tends to interact with those of low degree, which reveals the disassortativity and modularity of miRFNs. Our efforts in this study will be useful to further reveal the soybean miRNA-miRNA and miRNA-gene interactive mechanism on a systematic level. AVAILABILITY AND IMPLEMENTATION: A web tool for information retrieval and analysis of soybean miRFNs and the relevant target functional gene networks can be accessed at SoymiRNet: http://nclab.hit.edu.cn/SoymiRNet.

A novel insight into Gene Ontology semantic similarity.

Existing methods for computing the semantic similarity between Gene Ontology (GO) terms are often based on external datasets and, therefore are not intrinsic to GO. Furthermore, they not only fail to handle identical annotations but also show a strong bias toward well-annotated proteins when being used for measuring similarity of proteins. Inspired by the concept of cellular differentiation and dedifferentiation in developmental biology, we propose a shortest semantic differentiation distance (SSDD) based on the concept of semantic totipotency to measure the semantic similarity of GO terms and further compare the functional similarity of proteins. Using human ratings and a benchmark dataset, SSDD was found to improve upon existing methods for computing the semantic similarity of GO terms. An in-depth analysis shows that SSDD is able to distinguish identical annotations and does not depend on annotation richness, thus producing more unbiased and reliable results. Online services can be accessed at the Gene Functional Similarity Analysis Tools website (GFSAT: http://nclab.hit.edu.cn/GFSAT).

Single-cell transcriptome profiling implicates the psychological stress-induced disruption of spermatogenesis.

Male infertility has emerged as a global issue, partly attributed to psychological stress. However, the cellular and molecular mechanisms underlying the adverse effects of psychological stress on male reproductive function remain elusive. We created a psychologically stressed model using terrified-sound and profiled the testes from stressed and control rats using single-cell RNA sequencing. Comparative and comprehensive transcriptome analyses of 11,744 testicular cells depicted the cellular landscape of spermatogenesis and revealed significant molecular alterations of spermatogenesis suffering from psychological stress. At the cellular level, stressed rats exhibited delayed spermatogenesis at the spermatogonia and pachytene phases, resulting in reduced sperm production. Additionally, psychological stress rewired cellular interactions among germ cells, negatively impacting reproductive development. Molecularly, we observed the down-regulation of anti-oxidation-related genes and up-regulation of genes promoting reactive oxygen species (ROS) generation in the stress group. These alterations led to elevated ROS levels in testes, affecting the expression of key regulators such as ATF2 and STAR, which caused reproductive damage through apoptosis or inhibition of testosterone synthesis. Overall, our study aimed to uncover the cellular and molecular mechanisms by which psychological stress disrupts spermatogenesis, offering insights into the mechanisms of psychological stress-induced male infertility in other species and promises in potential therapeutic targets.

Immune-Associated Gene Signatures Serve as a Promising Biomarker of Immunotherapeutic Prognosis for Renal Clear Cell Carcinoma.

As the most common type of renal cell carcinoma (RCC), the renal clear cell carcinoma (ccRCC) is highly malignant and insensitive to chemotherapy or radiotherapy. Although systemic immunotherapies have been successfully applied to ccRCC in recent years, screening for patients who can benefit most from these therapies is still essential and challenging due to immunological heterogeneity of ccRCC patients. To this end, we implemented a series of deep investigation on the expression and clinic data of ccRCC from The Cancer Genome Atlas (TCGA) International Consortium for Cancer Genomics (ICGC). We identified a total of 946 immune-related genes that were differentially expressed. Among them, five independent genes, including SHC1, WNT5A, NRP1, TGFA, and IL4R, were significantly associated with survival and used to construct the immune-related prognostic differential gene signature (IRPDGs). Then the ccRCC patients were categorized into high-risk and low-risk subgroups based on the median risk score of the IRPDGs. IRPDGs subgroups displays distinct genomic and immunological characteristics. Known immunotherapy-related genes show different mutation burden, wherein the mutation rate of VHL was higher than 40% in the two IRPDGs subgroups, and SETD2 and BAP1 mutations differed most between two groups with higher frequency in the high-risk subgroup. Moreover, IRPDGs subgroups had different abundance in tumor-infiltrating immune cells (TIICs) with distinct immunotherapy efficacy. Plasma cells, regulatory cells (Tregs), follicular helper T cells (Tfh), and M0 macrophages were enriched in the high-risk group with a higher tumor immune dysfunction and rejection (TIDE) score. In contrast, the low-risk group had abundant M1 macrophages, mast cell resting and dendritic cell resting infiltrates with lower TIDE score and benefited more from immune checkpoint inhibitors (ICI) treatment. Compared with other biomarkers, such as TIDE and tumor inflammatory signatures (TIS), IRPDGs demonstrated to be a better biomarker for assessing the prognosis of ccRCC and the efficacy of ICI treatment with the promise in screening precise patients for specific immunotherapies.

Computational Characterizing Necroptosis Reveals Implications for Immune Infiltration and Immunotherapy of Hepatocellular Carcinoma.

Necroptosis is a programmed form of necrotic cell death in regulating cancer ontogenesis, progression, and tumor microenvironment (TME) and could drive tumor-infiltrating cells to release pro-inflammatory cytokines, incurring strong immune responses. Nowadays, there are few identified biomarkers applied in clinical immunotherapy, and it is increasingly recognized that high levels of tumor necroptosis could enhance the response to immunotherapy. However, comprehensive characterization of necroptosis associated with TME and immunotherapy in Hepatocellular carcinoma (HCC) remains unexplored. Here, we computationally characterized necroptosis landscape in HCC samples from TCGA and ICGA cohorts and stratified them into two necroptosis clusters (A or B) with significantly different characteristics in clinical prognosis, immune cell function, and TME-landscapes. Additionally, to further evaluate the necroptosis levels of each sample, we established a novel necroptosis-related gene score (NRGscore). We further investigated the TME, tumor mutational burden (TMB), clinical response to immunotherapy, and chemotherapeutic drug sensitivity of HCC subgroups stratified by the necroptosis landscapes. The NRGscore is robust and highly predictive of HCC clinical outcomes. Further analysis indicated that the high NRGscore group resembles the immune-inflamed phenotype while the low score group is analogous to the immune-exclusion or metabolism phenotype. Additionally, the high NRGscore group is more sensitive to immune checkpoint blockade-based immunotherapy, which was further validated using an external HCC cohort, metastatic melanoma cohort, and advanced urothelial cancer cohort. Besides, the NRGscore was demonstrated as a potential biomarker for chemotherapy, wherein the high NRGscore patients with more tumor stem cell composition could be more sensitive to Cisplatin, Doxorubicin, Paclitaxel-based chemotherapy, and Sorafenib therapy. Collectively, a comprehensive characterization of the necroptosis in HCC suggested its implications for predicting immune infiltration and response to immunotherapy of HCC, providing promising strategies for treatment.

Integrative analysis of DNA methylation and gene expression profiles identifies MIR4435-2HG as an oncogenic lncRNA for glioma progression.

Long noncoding RNAs (lncRNAs) have been shown to play an important role in tumor biogenesis and prognosis. The glioma is a grade classified cancer, however, we still lack the knowledge on their function during glioma progression. While previous studies have shown how lncRNAs regulate protein-coding gene epigenetically, it is still unclear how lncRNAs are regulated epigenetically. In this study, we firstly analyzed the RNA-seq data systematically across grades II, IV, and IV of glioma samples. We identified 60 lncRNAs that are significantly differentially expressed over disease progression (DElncRNA), including well-known PVT1, HOTAIR, H19 and rarely studied CARD8-AS, MIR4435-2HG. Secondly, by integrating HM450K methylation microarray data, we demonstrated that some of the lncRNAs are epigenetically regulated by methylation. Thirdly, we developed a DESeq2-GSEA-ceRNA-survival analysis strategy to investigate their functions. Particularly, MIR4435-2HG is highly expressed in high-grade glioma and may have an impact on EMT and TNFα signaling pathway by functioning as a miRNA sponge of miR-125a-5p and miR-125b-5p to increase the expression of CD44. Our results revealed the dynamic expression of lncRNAs in glioma progression and their epigenetic regulation mechanism.

Applications of Single-Cell Sequencing for Multiomics.

Single-cell sequencing interrogates the sequence or chromatin information from individual cells with advanced next-generation sequencing technologies. It provides a higher resolution of cellular differences and a better understanding of the underlying genetic and epigenetic mechanisms of an individual cell in the context of its survival and adaptation to microenvironment. However, it is more challenging to perform single-cell sequencing and downstream data analysis, owing to the minimal amount of starting materials, sample loss, and contamination. In addition, due to the picogram level of the amount of nucleic acids used, heavy amplification is often needed during sample preparation of single-cell sequencing, resulting in the uneven coverage, noise, and inaccurate quantification of sequencing data. All these unique properties raise challenges in and thus high demands for computational methods that specifically fit single-cell sequencing data. We here comprehensively survey the current strategies and challenges for multiple single-cell sequencing, including single-cell transcriptome, genome, and epigenome, beginning with a brief introduction to multiple sequencing techniques for single cells.

Alternative splicing links histone modifications to stem cell fate decision.

BACKGROUND: Understanding the embryonic stem cell (ESC) fate decision between self-renewal and proper differentiation is important for developmental biology and regenerative medicine. Attention has focused on mechanisms involving histone modifications, alternative pre-messenger RNA splicing, and cell-cycle progression. However, their intricate interrelations and joint contributions to ESC fate decision remain unclear. RESULTS: We analyze the transcriptomes and epigenomes of human ESC and five types of differentiated cells. We identify thousands of alternatively spliced exons and reveal their development and lineage-dependent characterizations. Several histone modifications show dynamic changes in alternatively spliced exons and three are strongly associated with 52.8% of alternative splicing events upon hESC differentiation. The histone modification-associated alternatively spliced genes predominantly function in G2/M phases and ATM/ATR-mediated DNA damage response pathway for cell differentiation, whereas other alternatively spliced genes are enriched in the G1 phase and pathways for self-renewal. These results imply a potential epigenetic mechanism by which some histone modifications contribute to ESC fate decision through the regulation of alternative splicing in specific pathways and cell-cycle genes. Supported by experimental validations and extended datasets from Roadmap/ENCODE projects, we exemplify this mechanism by a cell-cycle-related transcription factor, PBX1, which regulates the pluripotency regulatory network by binding to NANOG. We suggest that the isoform switch from PBX1a to PBX1b links H3K36me3 to hESC fate determination through the PSIP1/SRSF1 adaptor, which results in the exon skipping of PBX1. CONCLUSION: We reveal the mechanism by which alternative splicing links histone modifications to stem cell fate decision.

Genome-Wide Association and Mechanistic Studies Indicate That Immune Response Contributes to Alzheimer's Disease Development.

Alzheimer's disease (AD) is the most common cause of dementia. Although genome-wide association study (GWAS) have reported hundreds of single-nucleotide polymorphisms (SNPs) and genes linked to AD, the mechanisms about how these SNPs modulate the development of AD remain largely unknown. In this study, we performed GWAS for three traits in cerebrospinal fluid (CSF) and one clinical trait in the Alzheimer's Disease Neuroimaging Initiative (ADNI) cohort. Our analysis identified five most significant AD related SNPs (FDR < 0.05) within or proximal to APOE, APOC1, and TOMM40. One of the SNPs was co-inherited with APOE allele 4, which is the most important genetic risk factor for AD. Three of the five SNPs were located in promoter or enhancer regions, and transcription factor (TF) binding affinity calculations showed dramatic changes (| Log2FC| > 2) of three TFs (PLAG1, RREB1, and ZBTB33) for two motifs containing SNPs rs2075650 and rs157580. In addition, our GWAS showed that both rs2075650 and rs157580 were significantly associated with the poliovirus receptor-related 2 (PVRL2) gene (FDR < 0.25), which is involved in spreading of herpes simplex virus (HSV). The altered regulation of PVRL2 may increase the susceptibility AD patients to HSV and other virus infections of the brain. Our work suggests that AD is a type of immune disorder driven by viral or microbial infections of the brain during aging.

SoyFN: a knowledge database of soybean functional networks.

Many databases for soybean genomic analysis have been built and made publicly available, but few of them contain knowledge specifically targeting the omics-level gene-gene, gene-microRNA (miRNA) and miRNA-miRNA interactions. Here, we present SoyFN, a knowledge database of soybean functional gene networks and miRNA functional networks. SoyFN provides user-friendly interfaces to retrieve, visualize, analyze and download the functional networks of soybean genes and miRNAs. In addition, it incorporates much information about KEGG pathways, gene ontology annotations and 3'-UTR sequences as well as many useful tools including SoySearch, ID mapping, Genome Browser, eFP Browser and promoter motif scan. SoyFN is a schema-free database that can be accessed as a Web service from any modern programming language using a simple Hypertext Transfer Protocol call. The Web site is implemented in Java, JavaScript, PHP, HTML and Apache, with all major browsers supported. We anticipate that this database will be useful for members of research communities both in soybean experimental science and bioinformatics. Database URL: http://nclab.hit.edu.cn/SoyFN.